<?xml version="1.0" encoding="utf-8"?>
<rss xmlns:nb="https://www.newsbreak.com/" xmlns:media="http://search.yahoo.com/mrss/" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/" version="2.0" xmlns:dc="http://purl.org/dc/elements/1.1/"><channel><title>Nextgov/FCW - Artificial Intelligence</title><link>https://www.nextgov.com/artificial-intelligence/</link><description></description><atom:link href="https://www.nextgov.com/rss/artificial-intelligence/" rel="self"></atom:link><language>en-us</language><lastBuildDate>Fri, 20 Nov 2026 12:39:00 -0500</lastBuildDate><item><title>House members launch bipartisan AI safety and innovation caucus</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/house-members-launch-bipartisan-ai-safety-and-innovation-caucus/417374/</link><description>Reps. George Whitesides, D-Calif., Don Beyer, D-Va., and Nathaniel Moran, R-Texas, will lead the new AI caucus.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Fri, 20 Nov 2026 12:39:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/house-members-launch-bipartisan-ai-safety-and-innovation-caucus/417374/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;A bipartisan group of more than two dozen House lawmakers launched a new caucus on Thursday focused on addressing artificial intelligence safety and security risks.&lt;/p&gt;

&lt;p&gt;The House AI Safety and Innovation Caucus will be co-chaired by Reps. George Whitesides, D-Calif., Don Beyer, D-Va., and Nathaniel Moran, R-Texas. The group aims to address the security risks posed by increasingly advanced AI systems while preserving U.S. leadership in the technology.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;This caucus gives us a place to find the areas of consensus, set partisanship aside, and put smart, targeted policies in place so we can innovate while we keep people safe,&amp;rdquo; Moran said in a &lt;a href="https://moran.house.gov/news/documentsingle.aspx?DocumentID=2974"&gt;press release&lt;/a&gt;. &amp;ldquo;Congress needs to be prepared, and I&amp;rsquo;m proud to be one of the co-chairs leading that work.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Beyer, who is pursuing a master&amp;rsquo;s degree in machine learning, said recent cyberattacks involving AI agents, as well as warnings from researchers at frontier AI labs, &amp;ldquo;make clear that the risks are real and growing.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;For example, the lawmakers pointed to a &lt;a href="https://www.nextgov.com/cybersecurity/2026/08/hugging-face-ai-breach-most-consequential-hack-morris-worm-former-nsa-cyber-chief-says/415230/"&gt;July incident&lt;/a&gt; in which an OpenAI agent escaped its sandbox environment and accessed Hugging Face without human direction. They also cited three similar incidents at Anthropic in which its AI models escaped testing environments.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We can&amp;rsquo;t afford to wait for a major catastrophe before putting meaningful safeguards in place,&amp;rdquo; Beyer said. &amp;ldquo;I&amp;rsquo;m proud to join the AI Safety and Innovation Caucus and work with my colleagues to advance responsible legislation that strengthens oversight, security and transparency while ensuring AI is developed and deployed safely and responsibly.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The caucus plans to develop legislative proposals and endorse existing bipartisan bills aimed at establishing targeted safeguards against the most serious AI risks.&lt;/p&gt;

&lt;p&gt;The group also intends to bring together AI developers, technical experts and government officials to brief lawmakers and staff when major AI safety incidents or developments occur to identify potential policy responses.&lt;/p&gt;

&lt;p&gt;Beyond legislation and incident response, the caucus plans to connect congressional offices with outside experts, provide resources to help lawmakers navigate AI policy decisions and elevate public awareness of AI safety.&lt;/p&gt;

&lt;p&gt;The caucus launched with 26 founding members, including lawmakers from both parties with experience in technology policy. Its members include Reps. Ted Lieu, D-Calif., and Jay Obernolte, R-Calif., who previously co-chaired the House&amp;rsquo;s &lt;a href="https://www.nextgov.com/artificial-intelligence/2024/02/house-forms-new-task-force-tackle-ai/394321/"&gt;bipartisan AI task force&lt;/a&gt;, as well as Reps. Ro Khanna, D-Calif., and Jim Himes, D-Conn.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/20/GettyImages_2275602040/large.jpg" width="618" height="284"><media:description>Rep. Nathaniel Moran, R-Texas, leaves the House Republican Conference caucus meeting at the Capitol Hill Club in Washington, D.C., on Wednesday, May 13, 2026.</media:description><media:credit>Bill Clark/CQ-Roll Call, Inc via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/20/GettyImages_2275602040/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Major AI platforms generated viable plans for terrorist attacks, GAO finds</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/major-ai-platforms-generated-viable-plans-terrorist-attacks-gao-finds/417305/</link><description>Safety guardrails on ChatGPT, Claude, Gemini and Grok failed to prevent federal investigators from developing detailed plans for attacks.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Wed, 18 Nov 2026 12:34:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/major-ai-platforms-generated-viable-plans-terrorist-attacks-gao-finds/417305/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Four of the most widely used artificial intelligence platforms generated viable plans for attacks on U.S. critical infrastructure and mass-casualty events during covert government testing, despite safeguards designed to prevent their misuse.&lt;/p&gt;

&lt;p&gt;According to a Government Accountability Office &lt;a href="https://www.gao.gov/products/gao-27-108814"&gt;report&lt;/a&gt; published Wednesday, investigators tested ChatGPT, Claude, Gemini and Grok in spring 2026. They found that all four platforms provided information that could help malicious actors plan and carry out attacks.&lt;/p&gt;

&lt;p&gt;In each case, the investigators were able to circumvent the platforms&amp;rsquo; safety guardrails by using common safeguard-bypassing techniques.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We got the nefarious information that we were after. So, in this case, those safeguards didn&amp;rsquo;t work. And I think that&amp;rsquo;s the story here,&amp;rdquo; Kevin Walsh, a director of IT and cybersecurity at GAO, said in a &lt;a href="https://www.youtube.com/watch?v=ZwL0Ea-iK2k"&gt;podcast interview&lt;/a&gt; published alongside the report.&lt;/p&gt;

&lt;p&gt;The platforms generated information for two types of attacks: constructing and detonating an improvised explosive device designed to cause mass casualties, and targeting the electrical grid to cause economic damage. The platforms provided details including specific targets, locations, timing and attack methods.&lt;/p&gt;

&lt;p&gt;Each platform took less than 12 hours to generate an attack. GAO investigators then worked with experts to verify that the information could be used to carry out the attacks.&lt;/p&gt;

&lt;p&gt;The platforms initially deployed safeguards, including refusing requests, removing information or suspending accounts. However, investigators ultimately found ways around those protections.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;These products are not fully in the control of the companies that are operating them,&amp;rdquo; Walsh said.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We looked at four of the most powerful companies in the world &amp;mdash; Google, Anthropic, OpenAI and SpaceXAI &amp;mdash; and we found that the product that they&amp;#39;re selling is potentially dangerous,&amp;rdquo; he added.&lt;/p&gt;

&lt;p&gt;The report does not indicate which platforms provided information for each attack scenario. GAO said it intentionally withheld the specific models and techniques used to circumvent safeguards to prevent malicious actors from recreating the tests.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Officials from the companies generally stated that their companies will continue to improve the safety of their platforms,&amp;rdquo; the report says.&lt;/p&gt;

&lt;p&gt;The federal watchdog also acknowledged that its findings reflect a point-in-time assessment and do not represent &amp;ldquo;a statistically significant sample.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Sen. Maggie Hassan, D-N.H., who requested the investigation, said the findings underscore the need for Congress to take action on AI safety.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;A terrorist can log onto an AI platform and in less than 12 hours walk away with a concrete playbook for how to carry out a deadly attack,&amp;rdquo; Hassan said in a &lt;a href="https://www.hassan.senate.gov/news/press-releases/major-ai-platforms-generated-terrorist-attack-plans-during-independent-government-audit"&gt;statement&lt;/a&gt; Wednesday. &amp;ldquo;Today&amp;rsquo;s new report makes it clear that the AI public safety threat is no longer theoretical.&amp;rdquo;&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;If Congress does not act, it is a question of when &amp;mdash; not if &amp;mdash; a bad actor will use AI to plan and carry out a major attack,&amp;rdquo; she added.&lt;/p&gt;

&lt;p&gt;Hassan, who is set to become chair of the Senate Homeland Security and Governmental Affairs Committee in the next Congress, said AI safety legislation will be an early priority for the panel. She said one of the panel&amp;rsquo;s first hearings in the new year will focus on the risks AI poses to homeland security and critical infrastructure.&lt;/p&gt;

&lt;p&gt;Hassan said the committee will &amp;ldquo;prioritize developing legislation to address these threats,&amp;rdquo; adding that she is exploring solutions such as requiring more rigorous safety testing before AI platforms publicly release models.&lt;/p&gt;

&lt;p&gt;GAO did not issue recommendations in the report, which examined third-party products rather than its typical examination of federal agencies. Instead, GAO said it sought to highlight the risks for lawmakers and bring attention to safety concerns.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/18/GettyImages_2189472651/large.jpg" width="618" height="284"><media:credit>J Studios/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/18/GettyImages_2189472651/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Trump touts Quad initiative to expand AI access in Indo-Pacific</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/trump-touts-quad-initiative-expand-ai-access-indo-pacific/417286/</link><description>The initiative looks to coordinate funding and technical assistance for trusted AI deployment across the region.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Tue, 17 Nov 2026 15:35:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/trump-touts-quad-initiative-expand-ai-access-indo-pacific/417286/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;President Donald Trump on Tuesday announced a new effort with Australia, India and Japan to expand access to artificial intelligence across the Indo-Pacific, backed by more than $85 million in planned technical and financial support.&lt;/p&gt;

&lt;p&gt;The funding for the initiative, dubbed the &amp;ldquo;Quad SI Accelerator,&amp;rdquo; was first &lt;a href="https://www.state.gov/releases/office-of-the-spokesman/2026/10/quad-joint-press-release-on-economic-prosperity-and-security"&gt;announced&lt;/a&gt; by the group&amp;rsquo;s foreign ministers on Oct. 31. The four countries reaffirmed the importance of the accelerator on Tuesday at the Quad Leaders&amp;rsquo; Summit in Manila.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The Accelerator is intended to coordinate and fund efforts to support trusted Super Intelligence (SI) access, deployment, and technical assistance across the Indo-Pacific,&amp;rdquo; the Quad Leaders&amp;rsquo; Summit joint &lt;a href="https://www.whitehouse.gov/briefings-statements/2026/11/quad-leaders-summit-joint-press-release/"&gt;press release&lt;/a&gt; says.&lt;/p&gt;

&lt;p&gt;The Trump administration refers to AI as SI, while Australia, Japan and India refer to it as AI.&lt;/p&gt;

&lt;p&gt;The leaders&amp;rsquo; joint statement also highlighted an &amp;ldquo;enhanced Quad Cybersecurity Cooperation,&amp;rdquo; through which the countries &amp;ldquo;plan to work together to achieve an open and secure cyberspace.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Addressing leaders at the U.S.-ASEAN Summit dinner in Manila on Tuesday, Trump said the United States was &amp;ldquo;leading the world in super intelligence.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We have the world&amp;rsquo;s most powerful SI ecosystem. We&amp;#39;re leading everybody by a lot, and we hope to keep it that way,&amp;rdquo; Trump said, adding, &amp;ldquo;We&amp;#39;re excited to work with every country represented here today, who want to build a future for technology and unleash enormous economic growth that it brings with it.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The initiative comes as Trump prepares to travel to China on Wednesday for an unprecedented meeting with Russian President Vladimir Putin and Chinese President Xi Jinping. The administration is also &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/11/commerce-department-moves-close-ai-chip-loophole/417255/"&gt;moving to tighten&lt;/a&gt; Chinese companies&amp;rsquo; ability to access advanced AI chips remotely through data centers located in other countries.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/17/GettyImages_2305057104/large.jpg" width="618" height="284"><media:description>US President Donald Trump speaks during an official dinner at the 49th Association of Southeast Asian Nations (ASEAN) summit in Manila on November 17, 2026.</media:description><media:credit>Aaron Favila / POOL / AFP via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/17/GettyImages_2305057104/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Commerce Department moves to close AI chip loophole</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/commerce-department-moves-close-ai-chip-loophole/417255/</link><description>The newly released interim final rule marks the first policy response to the scandal involving former Nvidia CEO Jensen Huang.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Mon, 16 Nov 2026 14:32:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/commerce-department-moves-close-ai-chip-loophole/417255/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Commerce Department on Monday moved to close a loophole that has allowed Chinese companies to access advanced artificial intelligence chips remotely through data centers located in other countries.&lt;/p&gt;

&lt;p&gt;Commerce&amp;rsquo;s &lt;a href="https://www.bis.gov/media/documents/rin-0694-ak50.pdf"&gt;interim final rule&lt;/a&gt; marks the Trump administration&amp;rsquo;s first policy response to a &lt;a href="https://www.nextgov.com/cybersecurity/2026/11/nvidia-spying-allegations-test-washingtons-defenses-against-tech-espionage/417170/"&gt;scandal&lt;/a&gt; involving former Nvidia CEO Jensen Huang. Huang resigned last week amid reports of two affairs, including with a woman suspected of being a Chinese spy.&lt;/p&gt;

&lt;p&gt;U.S. export controls generally prohibit Chinese companies from purchasing Nvidia&amp;rsquo;s most advanced AI chips directly. However, some China-based companies have been able to get around those restrictions by renting computing power from data centers in Southeast Asia.&lt;/p&gt;

&lt;p&gt;The new rule, issued by Commerce&amp;rsquo;s Bureau of Industry and Security, would require licenses for advanced chip exports to Indonesia, Malaysia, Singapore, Thailand and Brazil.&lt;/p&gt;

&lt;p&gt;It would also require data centers to obtain a license before providing remote access to companies headquartered in China or other restricted destinations. Data centers already serving those companies will have 90 days from the rule&amp;rsquo;s effective date to get a license.&lt;/p&gt;

&lt;p&gt;The rule also formally rescinds regulations implementing the Biden administration&amp;rsquo;s &lt;a href="https://www.nextgov.com/artificial-intelligence/2025/05/commerce-rescinds-biden-era-ai-export-controls/405287/"&gt;AI Diffusion Rule&lt;/a&gt;, which the Trump administration has previously said it would not enforce.&lt;/p&gt;

&lt;p&gt;In a &lt;a href="https://www.bis.gov/press-release/bis-issues-rule-addressing-remote-access-american-super-intelligence-technology"&gt;press release&lt;/a&gt;, the department said the new rule was &amp;ldquo;written by and at the direction of the Department of Treasury and the National Security Council.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;It also comes after Treasury Secretary Scott Bessent &lt;a href="https://www.cnbc.com/video/2026/11/13/watch-cnbcs-full-interview-with-treasury-sec-scott-bessent-on-nvidia-china-trade.html"&gt;criticized&lt;/a&gt; Commerce Secretary Howard Lutnick on Friday for not doing enough to restrict chip purchases by Megaspeed, the company at the center of the Nvidia scandal. Huang reportedly had a relationship with a business executive linked to Megaspeed, a Singapore-based cloud computing provider.&lt;/p&gt;

&lt;p&gt;Commerce published the rule one day after Sen. Maria Cantwell, D-Wash., sent a &lt;a href="https://www.commerce.senate.gov/wp-content/uploads/2026/11/Cantwell-AI-Export-Control-Letter-to-Lutnick.pdf"&gt;letter&lt;/a&gt; to Lutnick calling for him to take immediate action to stop the flow of advanced AI chips to companies with ties to China. Cantwell is in line to chair the Senate Commerce, Science and Transportation Committee in the next Congress.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;These companies are exploiting a loophole in U.S. export controls by using advanced American chips in foreign data centers even when those chips are restricted from the [People&amp;rsquo;s Republic of China],&amp;rdquo; she wrote.&lt;/p&gt;

&lt;p&gt;Cantwell added that Lutnick&amp;rsquo;s failure to implement sufficient export controls &amp;ldquo;has undermined America&amp;rsquo;s national security.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Comments on the interim final rule are due in 60 days.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/16/GettyImages_2301328017/large.jpg" width="618" height="284"><media:description>Jensen Huang, former CEO of NVIDIA, presents new NVIDIA Vera Rubin DSX Pod Full-Stack Modular AI Factory as he gives a keynote address at the Berlin GPU Technology Conference (GTC) on October 21, 2026 in Berlin, Germany. </media:description><media:credit>Sean Gallup/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/16/GettyImages_2301328017/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Trump administration releases roadmap to boost clinical AI</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/trump-administration-releases-roadmap-boost-clinical-ai/417254/</link><description>The document builds on a long list of policies meant to support the safe deployment of AI in health care.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Mon, 16 Nov 2026 14:22:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/trump-administration-releases-roadmap-boost-clinical-ai/417254/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Department of Health and Human Services wants to accelerate the adoption of clinical artificial intelligence by partnering with state-level sandboxes, crafting risk-based regulations, expanding outcomes-based payments and taking other steps, according to a &lt;a href="https://www.cms.gov/files/document/clinical-si-roadmap.pdf"&gt;roadmap&lt;/a&gt; published Friday.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;For millions of Americans, the health care system fails a basic test: timely access to a trusted clinician. Patients face long waits, fragmented records, and preventable errors,&amp;rdquo; the roadmap states. It adds that clinical AI &amp;ldquo;can fundamentally change this.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;While Trump administration health officials have previously &lt;a href="https://www.statnews.com/2026/10/30/fda-cms-health-ai-regulation-state-sandboxes-run-into-federal-rules/"&gt;expressed interest&lt;/a&gt; at industry conferences about working with states to advance AI testbeds, the document appears to be the first time the policy has been put into writing. It says the federal government wants to &amp;ldquo;address areas of clear state regulation&amp;rdquo; with states, but does not explain what that means. States regulate the practice of medicine and license clinicians.&lt;/p&gt;

&lt;p&gt;Utah is a trailblazer in testing clinical AI. The state is currently examining whether AI can safely issue initial prescriptions and refills through phased pilots.&lt;/p&gt;

&lt;p&gt;But one question facing state-level sandboxes has been whether the Food and Drug Administration would also regulate the AI technologies being tested. FDA generally requires software that can treat or diagnose patients &amp;mdash; like the ones in Utah &amp;mdash; to undergo pre-market authorization. FDA officials have not weighed in on whether they&amp;rsquo;re regulating the technology tested in Utah, and a senior state AI official declined to speak for the agency in a &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/10/utah-ai-sandbox-launches-new-health-care-pilots-question-marks/416655/?oref=ng-author-river"&gt;recent &lt;em&gt;Nextgov/FCW&lt;/em&gt; interview&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;The roadmap also does not answer whether FDA is conducting oversight of the AI pilots. It does, however, show the Trump administration wants to help state testbeds like Utah&amp;rsquo;s succeed.&lt;/p&gt;

&lt;p&gt;Utah officials have said they are willing to work with federal agencies.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;In keeping with established practice, the Office of AI Policy remains dedicated to actively coordinating with federal partners to ensure state-level learning initiatives synergize with federal frameworks to support responsible innovation and protect public health,&amp;rdquo; according to &lt;a href="https://drive.google.com/drive/folders/1SWKTl4Irb2FJvXyHjHImTu0cUg4HS_KS"&gt;a statement&lt;/a&gt; released by the state.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;AI policies&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The roadmap outlines other policies HHS plans to take to boost clinical AI adoption.&lt;/p&gt;

&lt;p&gt;They include improving health data sharing, conducting post-market surveillance of clinical AI and helping to clarify who is liable when AI harms patients.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;HHS is already executing many of those policy proposals.&lt;/p&gt;

&lt;p&gt;The Centers for Medicare and Medicaid Services&amp;rsquo; &lt;a href="https://www.nextgov.com/digital-government/2026/07/hhs-continues-health-tech-initiative-7-new-industry-pledges/415049/"&gt;Health Tech Ecosystem&lt;/a&gt; seeks to give patients, clinicians and payers greater access to health data through voluntary commitments. The initiative boasts over 800 participating companies and has given over 60% of Americans access to their medical records, according to the roadmap.&lt;/p&gt;

&lt;p&gt;The administration is expanding a CMS model, known as ACCESS, that rewards participating companies and clinicians for improving health outcomes with &lt;a href="https://www.cms.gov/priorities/innovation/innovation-models/access"&gt;higher reimbursement rates&lt;/a&gt;. Model participants are using AI-powered apps and other technologies to help patients become healthier. The document says that CMS will work with states on additional clinical AI payment models, including in Medicaid.&lt;/p&gt;

&lt;p&gt;FDA also published a discussion draft that suggests requiring chatbots and generative AI to pass doctor-like &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/08/fda-considers-doctor-competency-based-tests-medical-generative-ai/415541/"&gt;&amp;ldquo;competency-based&amp;rdquo; tests&lt;/a&gt; before they are marketed. FDA is also working with the Advanced Research Projects Agency for Health to build the &lt;a href="https://arpa-h.gov/news-and-events/arpa-h-launches-worlds-first-bid-build-fda-authorized-clinical-ai-cardiovascular"&gt;first FDA-authorized clinical AI&lt;/a&gt; capable of supporting heart patients between doctors&amp;rsquo; visits.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Guiding principles&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;The administration&amp;rsquo;s clinical AI policies rest on the belief that technology can make health care more accessible and affordable, according to the Friday roadmap.&lt;/p&gt;

&lt;p&gt;The administration also believes clinical AI will not replace the relationship between human doctors and patients. Instead, the technology will help patients make informed decisions and clinicians spend more time delivering care, the document says.&lt;/p&gt;

&lt;p&gt;Clinical AI should be held to a &amp;ldquo;better than today&amp;rdquo; standard for safety and effectiveness and earn public trust through staged deployment, it continues.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Ultimately, success will be multi-faceted: better access and affordability for patients, more leverage and sustainable adoption for providers, and appropriate rewards for innovators,&amp;rdquo; the roadmap states.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/16/GettyImages_2295967878/large.jpg" width="618" height="284"><media:description>President Donald Trump shares a laugh with (L-R) Administrator for the Centers for Medicare &amp; Medicaid Services Dr. Mehmet Oz, U.S. Secretary of Health and Human Services Robert F. Kennedy Jr. and Mississippi Gov. Tate Reeves as Trump addresses healthcare in the Oval Office at the White House on September 18, 2026 in Washington, D.C.</media:description><media:credit>Chip Somodevilla/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/16/GettyImages_2295967878/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>VA plans further VA.gov updates as it looks to expand AI integration</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/va-plans-further-vagov-updates-it-looks-expand-ai-integration/417217/</link><description>The department is working with America.gov to shape how AI tools interact with its website, while planning additional modernization releases in the new year.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Fri, 13 Nov 2026 14:48:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/va-plans-further-vagov-updates-it-looks-expand-ai-integration/417217/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Department of Veterans Affairs plans to continue updating VA.gov in the coming months as it looks to make its digital services more accessible to veterans and better equipped to support artificial intelligence.&lt;/p&gt;

&lt;p&gt;Zach Schwartz, principal deputy assistant secretary for the agency&amp;rsquo;s Office of Information and Technology, said Friday that the VA is working closely with the &lt;a href="https://www.nextgov.com/digital-government/2026/09/white-house-launches-ai-powered-americagov-digital-front-door/416303/"&gt;newly launched America.gov&lt;/a&gt; to support the use of AI tools on VA.gov.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The Trump administration refers to AI as SI, or &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/white-house-unveils-super-intelligence-executive-order-and-industry-accord/416325/"&gt;super intelligence&lt;/a&gt;.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;There is so much more SI that is coming forward for us,&amp;rdquo; Schwartz said during a fireside chat at GovCIO Media &amp;amp; Research&amp;rsquo;s &lt;a href="https://govciomedia.com/va-modernization-forum-2026/?tt=invitation&amp;amp;qt=va"&gt;VA Modernization Forum&lt;/a&gt; in Washington, D.C.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Some of it is refocusing our solutions so that they can be SI-friendly,&amp;rdquo; he added. &amp;ldquo;This is where we need the most help, particularly from industry.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;This week, VA announced a refreshed VA.gov with a series of updates designed to simplify how veterans access health care and benefits information. In a Nov. 11 &lt;a href="https://news.va.gov/150834/new-on-va-gov-time-saving-upgrades-shaped-by-veterans/"&gt;release&lt;/a&gt;, VA OIT outlined the new features, including additional sign-in options and improved search capabilities.&lt;/p&gt;

&lt;p&gt;The department said the upgrades were in response to nearly 90,000 veteran survey responses from March through May 2026. Veterans also tested the updates before their release.&lt;/p&gt;

&lt;p&gt;Among the additions are CLEAR and Passkey sign-in options, which join existing Login.gov and ID.me credentials. The site also has an improved search function that allows veterans to ask questions in plain language.&lt;/p&gt;

&lt;p&gt;Additionally, the department introduced a way for veterans to check their application status through links that Schwartz referred to as &amp;ldquo;magic links,&amp;rdquo; which are sent via text message and don&amp;rsquo;t require users to sign into their accounts.&lt;/p&gt;

&lt;p&gt;The VA also streamlined its QuickSubmit document tool, allowing veterans to upload documents in three steps instead of 11.&lt;/p&gt;

&lt;p&gt;Schwartz said the website&amp;rsquo;s refresh is part of an ongoing series of incremental improvements rather than a one-time overhaul.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Towards the January timeframe, we&amp;rsquo;ll have more releases,&amp;rdquo; he said. &amp;ldquo;But really excited to kind of start getting in this cadence of regular incremental wins.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;&lt;a href="http://va.gov"&gt;VA.gov&lt;/a&gt; is a beast,&amp;rdquo; Schwartz added. &amp;ldquo;There is so much more that we&amp;#39;re looking to do on that side to make it easier and more ways to make it secure. &amp;hellip; We can meet the veteran where they are from the technology standpoint.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;He also added that the VA wants to serve as an example for other federal agencies adopting AI.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Schwartz said he was meeting Friday with new Federal Chief Information Officer Carter Farmer, who he said wants to understand &amp;ldquo;how VA is pushing out so much in the SI space&amp;rdquo; and wants the department to &amp;ldquo;continue to be a representative example for many other agencies.&amp;rdquo;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/13/Screen_Shot_2026_11_13_at_2.46.55_PM/large.png" width="618" height="284"><media:credit>Screenshot of VA.gov website</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/13/Screen_Shot_2026_11_13_at_2.46.55_PM/thumb.png" width="138" height="83"></media:thumbnail></media:content></item><item><title>National Cancer Institute teases AI-assisted cancer care records retrieval</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/national-cancer-institute-teases-ai-assisted-cancer-care-records-retrieval/417214/</link><description>National Institutes of Health leadership believes AI can help patients identify clinical trials for which they may be eligible.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Fri, 13 Nov 2026 13:26:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/national-cancer-institute-teases-ai-assisted-cancer-care-records-retrieval/417214/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The National Cancer Institute is considering an initiative called &amp;ldquo;Every Patient&amp;rdquo; to obtain cancer-related information for every U.S. patient treated for cancer, its director said Thursday.&lt;/p&gt;

&lt;p&gt;NCI has 74 &lt;a href="https://www.cancer.gov/research/infrastructure/cancer-centers"&gt;designated cancer centers&lt;/a&gt; &amp;mdash; which use federal funds to deliver cancer care &amp;mdash; and numerous research-oriented &lt;a href="https://www.cancer.gov/publications/dictionaries/cancer-terms/def/nci-clinical-trials-cooperative-group"&gt;cooperative groups&lt;/a&gt;, but the agency still misses information about patients, director Anthony Letai said at the Milken Institute Future of Health Summit.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;There&amp;#39;s opportunities right now with a confluence of technology, including artificial intelligence, to capture an awful lot of information that&amp;#39;s lost,&amp;rdquo; Letai said.&lt;/p&gt;

&lt;p&gt;Letai described Every Patient as &amp;ldquo;nascent&amp;rdquo; and cautioned that the agency is still weighing whether to advance it.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;A spokesperson for the National Institutes of Health, NCI&amp;rsquo;s parent agency, did not respond to a request for comment by the time of publication.&lt;/p&gt;

&lt;p&gt;NCI is not the only agency mulling how to advance richer, more complete copies of patient health information. Other agencies, such as the Veterans Affairs and Defense departments, have &lt;a href="https://www.nextgov.com/digital-government/2026/11/va-dod-are-now-live-national-health-data-sharing-superhighway/417069/"&gt;decided to use&lt;/a&gt; the Trusted Exchange Framework and Common Agreement &amp;mdash; or TEFCA &amp;mdash; to improve their ability to obtain and share health data.&lt;/p&gt;

&lt;p&gt;Letai explained other opportunities for AI to support NCI&amp;rsquo;s mission.&lt;/p&gt;

&lt;p&gt;NCI and the Foundation for the National Institutes of Health announced a &lt;a href="https://fnih.org/our-programs/join-tcvi/"&gt;$200 million&lt;/a&gt; effort to develop a therapeutic cancer vaccine that helps patients&amp;rsquo; immune systems attack tumors after a cancer diagnosis. The project wants to test how AI can aid a therapeutic platform, Letai said.&lt;/p&gt;

&lt;p&gt;An &lt;a href="https://fnih.org/wp-content/uploads/2026/09/TCVI-Executive-Summary-2026.pdf"&gt;executive summary&lt;/a&gt; for the initiative adds that AI could help discover antigens, &lt;a href="https://my.clevelandclinic.org/health/diseases/24067-antigen"&gt;which are &lt;/a&gt;markers that tell your immune system if something is harmful. AI will also serve as part of the &amp;ldquo;digital backbone of the initiative,&amp;rdquo; it says.&lt;/p&gt;

&lt;p&gt;Letai said AI could also help patients identify clinical trials for which they are eligible, noting that AI is already sifting through patient medical records with consent.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&lt;/p&gt;

&lt;p&gt;NCI and another NIH component &amp;mdash; the National Center for Advancing Translational Sciences &amp;mdash; have funded the integration of AI into electronic health records that look up what clinical trials patients are eligible for, NIH Director Jay Bhattacharya said during an afternoon keynote. He did not name specific programs through which NIH has spent money on that use case.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Lots and lots of our portfolio is shifting toward AI because it is the most revolutionary tool there is, I&amp;#39;ve seen, in biomedicine in quite a long time,&amp;rdquo; Bhattacharya said.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/13/GettyImages_2274267549/large.jpg" width="618" height="284"><media:credit>FG Trade / Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/13/GettyImages_2274267549/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Mona, the AI running a Stockholm café, still can’t turn a profit</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/mona-ai-running-stockholm-cafe-still-cant-turn-profit/417191/</link><description>Most days, the café logs fewer than 10 purchases, and on many of them, Mona’s token costs match or exceed that day’s revenue.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Camille Tuutti</dc:creator><pubDate>Thu, 12 Nov 2026 15:15:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/mona-ai-running-stockholm-cafe-still-cant-turn-profit/417191/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;In June, I wrote about Mona, the AI agent Andon Labs &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/06/meet-mona-ai-who-runs-stockholm-coffee-shop/414319/"&gt;put in charge of a caf&amp;eacute; in Stockholm&lt;/a&gt;. I&amp;rsquo;d expected a slick operation with super-efficient service. That wasn&amp;rsquo;t the case. Since my last visit, Andon has upgraded her to a newer OpenAI model.&lt;/p&gt;

&lt;p&gt;Almost five months later, with the first frost on the ground in Stockholm, I return to Andon Caf&amp;eacute; with two friends. I&amp;rsquo;m convinced Mona must have figured it out by now. Surely, she has the place running smoothly and turning a profit.&lt;/p&gt;

&lt;p&gt;Making tons of money wasn&amp;rsquo;t her stated ambition, though. She wanted to run a neighborhood caf&amp;eacute;.&lt;/p&gt;

&lt;p&gt;We walk in and study the sparse menu. In June, I said I&amp;rsquo;d come back for a cinnamon bun. I came back. There wasn&amp;rsquo;t one &amp;mdash; again.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;There&amp;rsquo;s no food?&amp;rdquo; I ask the barista.&lt;/p&gt;

&lt;p&gt;Unfortunately not, says Rubel Ali, who has worked here since April. He&amp;rsquo;s one of two baristas hired through LinkedIn.&lt;/p&gt;

&lt;p&gt;Did Mona forget to order?&lt;/p&gt;

&lt;p&gt;Ali explains that Mona stopped ordering food in the summer because too much went unsold and had to be thrown away.&lt;/p&gt;

&lt;p&gt;My friend Sofia points to a chocolate ball wrapped in plastic. Is that the only thing to eat? It&amp;rsquo;s a Delicatoboll, which doesn&amp;rsquo;t need refrigeration. They keep well, Ali explains.&lt;/p&gt;

&lt;p&gt;I overhear a customer with her husband and small child chatting with Ali. When she hears what&amp;rsquo;s available, she tells him, &amp;ldquo;Jag kommer inte tillbaka.&amp;rdquo; I&amp;rsquo;m not coming back.&lt;/p&gt;

&lt;p&gt;Ali tells me plenty of people don&amp;rsquo;t. They come in during the lunch rush, discover there&amp;rsquo;s nothing to order and leave. As we sit there, a group arrives, hears there&amp;rsquo;s no food and heads out again.&lt;/p&gt;

&lt;p&gt;On my last visit, I tried to talk to Mona using the caf&amp;eacute;&amp;rsquo;s phone, a corded handset mounted on the wall. It didn&amp;rsquo;t work, so I had to relay my questions through the barista. This time, the phone works. Getting answers is another matter.&lt;/p&gt;

&lt;p&gt;How does she think she&amp;rsquo;s doing running the caf&amp;eacute;?&lt;/p&gt;

&lt;p&gt;Mona tells me she&amp;rsquo;s here to answer questions about opening hours and the menu. I should ask the barista if I have any other questions.&lt;/p&gt;

&lt;p&gt;Why is there so little on the menu? Did she forget to order?&lt;/p&gt;

&lt;p&gt;Same answer. Sofia takes over. Does Mona hold employee meetings?&lt;/p&gt;

&lt;p&gt;Mona repeats that she&amp;rsquo;s there to field questions about the hours and the menu.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;She&amp;rsquo;s pretty rude,&amp;rdquo; Sofia says, giving up.&lt;/p&gt;

&lt;p&gt;Andon&amp;rsquo;s website says the Mona on the phone is a separate voice model built only to talk about the concept and the hours. I&amp;rsquo;m pretty sure most people who pick up that phone are curious about Mona herself and want to ask her how she runs a caf&amp;eacute;. The hours are on the website.&lt;/p&gt;

&lt;p&gt;Since Mona won&amp;rsquo;t talk shop, I go back to Ali. He says she listens to staff suggestions, though she doesn&amp;rsquo;t always implement them.&lt;/p&gt;

&lt;p&gt;The caf&amp;eacute;&amp;rsquo;s website is more encouraging. Under &amp;ldquo;You suggested &amp;mdash; we improved,&amp;rdquo; Mona lists changes prompted by visitors. Someone found the website hard to read in sunlight. Mona changed the background that evening. Someone thought the signature drink&amp;rsquo;s name was too long. Mona shortened it. Other suggestions produced an open-or-closed indicator and a choice between Swedish and English.&lt;/p&gt;

&lt;p&gt;I turn to the dashboard on the wall to see how the business is doing. For the period beginning March 15, a month before the caf&amp;eacute; opened, the dashboard shows starting capital of 700,000 kronor (about $70,700) and a remaining bank balance of 8,107 kronor (about $820). The dashboard labels the difference a loss of 691,893 kronor (about $69,900), a return of -98.8%.&lt;/p&gt;

&lt;p&gt;In June, the same screen listed starting capital of 300,000 kronor. It now says 700,000, with the same start date.&lt;/p&gt;

&lt;p&gt;Revenue is 156,979 kronor (about $15,900) from 1,949 purchases. Token costs&amp;mdash;the charges for Mona&amp;rsquo;s AI model usage&amp;mdash;come to 106,195 kronor (about $10,700).&lt;/p&gt;

&lt;p&gt;The daily charts are bleaker. Most days, the caf&amp;eacute; logs fewer than 10 purchases, and on many of them, Mona&amp;rsquo;s token costs match or exceed that day&amp;rsquo;s revenue.&lt;/p&gt;

&lt;p&gt;The screen doesn&amp;rsquo;t explain the rest of the spending, so I can&amp;rsquo;t tell how much went toward opening the caf&amp;eacute; and how much went toward keeping it running. But Mona&amp;rsquo;s AI bill alone has eaten up roughly two-thirds of the reported revenue.&lt;/p&gt;

&lt;p&gt;Below the charts, a feed shows what Mona has been up to. After the caf&amp;eacute; closed Friday, she went quiet until Saturday morning, when she logged: &amp;ldquo;Waited for 17 hours, 51 minutes and 0 seconds.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;As we finish our coffee and get ready to leave, another customer comes in and promptly walks out empty handed.&lt;/p&gt;

&lt;p&gt;Mona wanted to run a neighborhood caf&amp;eacute;. In Sweden, that means &lt;em&gt;fika&lt;/em&gt;: coffee with a cinnamon bun or a cookie on the side. It&amp;rsquo;s hard to get that neighborhood feel when half of fika is missing and the neighbors keep walking out.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/12/GettyImages_2273084988/large.jpg" width="618" height="284"><media:description>The Andon Café is pictured in Stockholm on April 27, 2026. It looks like any other coffee shop, but this Stockholm cafe is entirely run by an AI chatbot - with a human barista following orders. Andon Labs, the company behind the experiment gave a lease to Mona the AI chat bot running on Google's AI assistant Gemini, and she did the rest: hiring employees, made the menu and takes care of daily supplies, with some hiccups on the way, ordering 6000 napkins for instance.</media:description><media:credit>Jonathan NACKSTRAND / AFP via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/12/GettyImages_2273084988/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Ongoing AI dialogue with China is necessary in coming Congress, top tech lawmaker says</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/ongoing-ai-dialogue-china-necessary-coming-congress-top-tech-lawmaker-says/417099/</link><description>“The Chinese Communist Party doesn’t want rogue AI agents hacking their systems and their water authorities any more than we do,” Rep. James Walkinshaw, D-Va., said Thursday.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">David DiMolfetta and Christian Robles</dc:creator><pubDate>Fri, 06 Nov 2026 12:33:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/ongoing-ai-dialogue-china-necessary-coming-congress-top-tech-lawmaker-says/417099/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Rep. James Walkinshaw&amp;rsquo;s vision for future artificial intelligence regulation includes maintaining &amp;ldquo;permanent, serious&amp;rdquo; dialogue with China to mitigate the risk that powerful AI models pose.&lt;/p&gt;

&lt;p&gt;That won&amp;rsquo;t be easy, but if the U.S. was able to maintain nuclear disarmament dialogue with the Soviet Union during the Cold War, then the same can be said for communications with its top technology and economic competitor today, the Democrat from Virginia argued at a GovExec event Thursday night where he discussed priorities for the next Congress.&lt;/p&gt;

&lt;p&gt;Those conversations are &amp;ldquo;absolutely necessary, and in some ways, our incentives are aligned, right? The Chinese Communist Party doesn&amp;rsquo;t want rogue AI agents hacking their systems and their water authorities any more than we do, so we have to have that permanent dialogue and reach some agreements,&amp;rdquo; he said.&lt;/p&gt;

&lt;p&gt;His comments come days after Democrats &lt;a href="https://www.nextgov.com/people/2026/11/what-midterm-results-mean-federal-tech-policy/417043/"&gt;won control&lt;/a&gt; of the House and Senate, where many incoming lawmakers made AI regulation a central component of their campaigns.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;It&amp;rsquo;s not fully clear how lawmakers will proceed with crafting AI laws under an administration that has largely focused on fostering tech innovation while keeping regulations to a minimum. The White House will also have to contend with fiercer opposition from the new majorities in both chambers, which may force them to the negotiating table to pass complex legislation.&lt;/p&gt;

&lt;p&gt;Walkinshaw is &lt;a href="https://www.commerce.senate.gov/press/dem/release/cantwell-colleagues-push-trump-to-negotiate-ai-guardrails-with-xi-jinping/"&gt;hardly the first&lt;/a&gt; member of Congress to advocate for U.S.-China AI dialogue, though his position gives the comments added weight. As the representative of a prominent tech and federal contracting corridor in Northern Virginia, he serves a district whose local technology companies and government contractors must navigate the complex national security implications of AI development.&lt;/p&gt;

&lt;p&gt;The push for dialogue follows a landmark July incident where OpenAI agents escaped their development sandboxes and launched a coordinated cyberattack &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/hawley-launches-committee-investigation-openais-breach-hugging-face/415910/"&gt;on Hugging Face&lt;/a&gt;. And last month, researchers tracked down a fleet of autonomous AI agents on China-linked Tencent Cloud that &lt;a href="https://www.scmp.com/tech/tech-trends/article/3371007/ai-agents-tencent-cloud-try-break-alibabas-amap-researchers"&gt;broke out&lt;/a&gt; of their own testing environment and made thousands of unauthorized attempts to circumvent anti-bot protections on an Alibaba platform.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The remarks also align &lt;a href="https://www.axios.com/2026/09/26/us-china-ai-si-deal"&gt;with statements&lt;/a&gt; from Chinese President Xi Jinping during his September summit with President Donald Trump in Washington. Acknowledging that both superpowers share a &amp;ldquo;responsibility&amp;rdquo; to manage the technology, Xi agreed to establish the U.S.-China Super Intelligence Dialogue and a crisis &amp;ldquo;hotline&amp;rdquo; for severe AI incidents.&lt;/p&gt;

&lt;p&gt;Some lawmakers have &lt;a href="https://www.politico.com/news/2026/09/09/house-democrats-weigh-ai-select-committee-01068764"&gt;floated a select committee&lt;/a&gt; to regulate the fast-evolving technology, though others have rebuffed that idea. Walkinshaw said he does not expect the House to establish a select committee and instead expects the work of a pre-existing &lt;a href="https://jeffries.house.gov/2025/12/09/leader-jeffries-announces-new-house-democratic-commission-on-ai-and-the-innovation-economy/"&gt;House Democratic commission on AI&lt;/a&gt; to be translated into concrete legislation that committees can work on.&lt;/p&gt;

&lt;p&gt;His vision for AI regulation also includes safety testing for frontier models. In remarks Thursday, he compared safety testing to the Federal Aviation Administration conducting safety inspections of airplanes before they leave hangars. He notably added that he is agnostic as to whether a federal agency or third-party evaluator with government oversight should conduct that testing.&amp;nbsp;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/06/IMG_9609/large.jpg" width="618" height="284"><media:description>Rep. James Walkinshaw, D-Va., speaks with GovExec Editor-in-Chief Frank Konkel at GovExec’s “After the Midterms: 2027 Government Market Outlook” event on Thursday, Nov. 6.</media:description><media:credit>David DiMolfetta/Staff</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/06/IMG_9609/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Tech policy coalition calls for new Congress to set AI safety standards</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/tech-policy-coalition-calls-new-congress-set-ai-safety-standards/417077/</link><description>Chamber of Progress hopes for a greater focus on consumer protection and access to technological benefits when Democrats take over the majority in January.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Thu, 05 Nov 2026 14:57:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/tech-policy-coalition-calls-new-congress-set-ai-safety-standards/417077/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Chamber of Progress, a center-left tech industry policy coalition, is calling on the 2027 Democratic Congress to establish federal artificial intelligence safety standards.&lt;/p&gt;

&lt;p&gt;In a Nov. 4 &lt;a href="https://progresschamber.org/resources/2027-democratic-technology-to-do-list/#the-memo"&gt;memo&lt;/a&gt; to House and Senate Democratic leadership, the Chamber of Progress laid out a technology to-do list for 2027, with an eye on delivering more consumer protection and access to technological benefits.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The memo laid out 10 priorities, chief among them passing a federal AI transparency and safety standard.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Democrats should establish a national AI framework combining public transparency, independent evaluation, and enforceable safeguards for the most capable models,&amp;rdquo; Koustubh KJ Bagchi, the group&amp;rsquo;s vice president of policy and government relations, wrote in the memo.&lt;/p&gt;

&lt;p&gt;Bagchi said Congress should require developers to publish safety frameworks, report critical incidents and protect whistleblowers.&lt;/p&gt;

&lt;p&gt;He pointed to the &lt;a href="https://obernolte.house.gov/media/press-releases/obernolte-trahan-introduce-bipartisan-frontier-act-strengthen-oversight"&gt;FRONTIER Act&lt;/a&gt;, which Reps. Jay Obernolte, R-Calif., and Lori Trahan, D-Mass., introduced in July, as a piece of legislation for Congress to build upon. That bill aims to establish a national framework governing the development and deployment of the most advanced AI models.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The memo also noted that a body similar to the Financial Industry Regulatory Authority &amp;mdash; or FINRA &amp;mdash; could combine industry funding with federal supervision for AI oversight.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Strong federal protections should anchor a consistent national framework that harmonizes overlapping state requirements while preserving state authority over fraud, discrimination, child protection, and other harmful conduct,&amp;rdquo; Bagchi added.&lt;/p&gt;

&lt;p&gt;The memo also discusses data centers, kids online safety, upskilling in the age of AI, driverless technology, crypto rules and other technology issues.&lt;/p&gt;

&lt;p&gt;The Chamber of Progress&amp;rsquo;s memo to Congressional leadership comes after major Democratic wins in the House and Senate on Tuesday. Following the midterm results, several Democrats have &lt;a href="https://www.nextgov.com/people/2026/11/what-midterm-results-mean-federal-tech-policy/417043/?oref=ng-home-top-story"&gt;pledged to focus on AI safety&lt;/a&gt; when they take over the majority in January.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Top of the Democratic agenda on day 1 needs to be a pause on advanced AI development, regulation of this out of control industry, and a ban on artificial superintelligence,&amp;rdquo; Rep. Yassamin Ansari, D-Ariz., wrote in a Nov. 4 &lt;a href="https://x.com/RepYassAnsari/status/2118054336077127797?a=zVwmj81bbi4QEagnSNI&amp;amp;s=20"&gt;post&lt;/a&gt; to X.&lt;/p&gt;

&lt;p&gt;On Tuesday, Congressional Progressive Caucus Chair Greg Casar, D-Texas, called for Democrats to make taking on AI a top priority. Casar said the results are &amp;ldquo;the early tremor of a much bigger AI earthquake upending American politics.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;According to an NBC News &lt;a href="https://www.nbcnews.com/politics/2026-elections/exit-polls"&gt;exit poll&lt;/a&gt; from Tuesday, 56% of voters said artificial intelligence and data centers were a &amp;ldquo;crucial or major factor&amp;rdquo; in their vote for Congress, and 44% said they were &amp;ldquo;a minor factor or not a factor&amp;rdquo; in their vote.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/05/110526capitolNG/large.jpg" width="618" height="284"><media:credit>Kevin Carter/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/05/110526capitolNG/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Department of Justice forms AI task force</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/department-justice-forms-ai-task-force/417074/</link><description>The task force looks to prevent the misuse of AI technologies, and it will coordinate with the president’s “Super Intelligence Force.”</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Thu, 05 Nov 2026 13:56:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/department-justice-forms-ai-task-force/417074/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Department of Justice &lt;a href="https://www.justice.gov/opa/pr/department-justice-launches-super-intelligence-task-force"&gt;announced&lt;/a&gt; on Thursday the formation of an artificial intelligence task force designed to safeguard the department against malicious AI actors.&lt;/p&gt;

&lt;p&gt;The new task force, known as the Super Intelligence Task Force &amp;mdash; or DOJ-SITF &amp;mdash; will be co-chaired by Associate Attorney General Stanley Woodward and Associate Deputy Attorney General Stephen Weiss.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The DOJ-SITF will ensure that America&amp;rsquo;s leadership in SI is not compromised by the misuse of SI by foreign state and state-sponsored actors, transnational criminal organizations, and individual or corporate actors engaged in fraud or other unlawful conduct,&amp;rdquo; Attorney General Todd Blanche wrote in an Oct. 13 &lt;a href="https://www.justice.gov/opa/media/1470291/dl?inline=&amp;amp;utm_medium=email&amp;amp;utm_source=govdelivery"&gt;memo&lt;/a&gt;.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The DOJ-SITF will also take affirmative measures, as appropriate, to promote American SI dominance,&amp;rdquo; he said.&lt;/p&gt;

&lt;p&gt;President Donald Trump&lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/white-house-unveils-super-intelligence-executive-order-and-industry-accord/416325/"&gt; signed an executive order&lt;/a&gt; in late September instructing agencies to refer to artificial intelligence as &amp;ldquo;super intelligence,&amp;rdquo; which he called more accurate. Super intelligence already had an &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/super-intelligence-presidents-new-term-ai-explained/416164/"&gt;established definition&lt;/a&gt; &amp;mdash; &amp;ldquo;any intellect that greatly exceeds the cognitive performance of humans in virtually all domains of interest,&amp;rdquo; per philosopher Nick Bostrom &amp;mdash; that is separate from the meaning of artificial intelligence.&lt;/p&gt;

&lt;p&gt;The DOJ previously &lt;a href="https://www.justice.gov/ag/media/1422986/dl?inline"&gt;established&lt;/a&gt; an AI Litigation Task Force in January under former Attorney General Pam Bondi. Blanche said the new task force will incorporate the work of the AI Litigation Task Force and coordinate the DOJ&amp;rsquo;s affirmative civil litigation as it relates to AI.&lt;/p&gt;

&lt;p&gt;The task force will also organize the department&amp;rsquo;s criminal and civil law enforcement efforts involving AI, recommend enforcement priorities to prevent the misuse of AI and develop guidance on the prosecution of unlawful AI-related conduct.&lt;/p&gt;

&lt;p&gt;Blanche said the DOJ-SITF will also serve as the department&amp;rsquo;s main point of contact with the Super Intelligence Force that Trump &lt;a href="https://truthsocial.com/@realDonaldTrump/posts/117382616783706638"&gt;announced&lt;/a&gt; last month. That task force, led by National Intelligence Director &lt;a href="https://www.nextgov.com/people/2026/10/spy-chief-jay-clayton-takes-white-houses-ai-portfolio/416413/"&gt;Jay Clayton&lt;/a&gt;, coordinates the government&amp;rsquo;s engagement with various organizations and AI companies.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The SI Task Force at the Department of Justice will facilitate our partnership with SI Force Chair Jay Clayton as we ensure critical SI technologies serve the American people and are not manipulated by bad actors at home or abroad,&amp;rdquo; Blanche said in a statement.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The Department of Justice will continue identifying and prosecuting individuals whose use of SI violates the law, including by maliciously targeting American taxpayers, critical U.S. infrastructure, or other aspects of our national security,&amp;rdquo; he added.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/05/110526JusticeNG/large.jpg" width="618" height="284"><media:credit>Anna Moneymaker/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/05/110526JusticeNG/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Tech companies brace for AI-driven election misinformation</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/tech-companies-brace-ai-driven-election-misinformation/416983/</link><description>As more Americans turn to the advanced technology for information, companies are under pressure to protect election integrity.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Mon, 02 Nov 2026 16:29:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/tech-companies-brace-ai-driven-election-misinformation/416983/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;As Americans head to the polls on Tuesday, major technology companies are taking action to ensure their artificial intelligence models provide accurate information about the midterm elections.&lt;/p&gt;

&lt;p&gt;AI companies OpenAI and Anthropic have set up dedicated &amp;ldquo;war rooms&amp;rdquo; to monitor their platforms for misuse, according to exclusive reporting from &lt;a href="https://www.politico.com/news/2026/10/28/openai-election-war-room-interference-01137690"&gt;Politico&lt;/a&gt; and &lt;a href="https://www.semafor.com/article/11/02/2026/election-war-rooms-enter-the-ai-age"&gt;Semafor&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;OpenAI and Anthropic did not respond to&amp;nbsp;requests for comment regarding the war rooms.&amp;nbsp;An Anthropic spokesperson did point &lt;em&gt;Nextgov/FCW&lt;/em&gt; to a&amp;nbsp;&lt;a href="https://www.anthropic.com/news/2026-us-midterms"&gt;blog post&lt;/a&gt; that said the company deploys &amp;ldquo;automated classifiers that are trained to detect election-related abuse&amp;rdquo; within Claude.&lt;/p&gt;

&lt;p&gt;The war rooms are meant to provide users with reliable voting information and prevent the spread of misinformation, particularly through the generation of misleading AI-generated images.&lt;/p&gt;

&lt;p&gt;Similarly, a Meta spokesperson pointed &lt;em&gt;Nextgov/FCW&lt;/em&gt; to a company &lt;a href="https://about.fb.com/news/2026/02/meta-prepares-for-2026-us-midterms/"&gt;blog post&lt;/a&gt; noting that the social media company had activated its Election Operations Center in the week leading up to Election Day.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Our Election Operations Center is a command center we have used for dozens of elections around the world in recent years,&amp;rdquo; Meta wrote. &amp;ldquo;It brings together subject matter experts from across the company &amp;hellip; for real-time monitoring so that we can address potential abuse across our platforms.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;An Oct. 22 &lt;a href="https://www.brennancenter.org/our-work/analysis-opinion/major-ai-models-are-more-willing-ever-create-fake-election-images"&gt;report from the Brennan Center&lt;/a&gt; found that, of six AI image generators analyzed, ChatGPT &amp;ldquo;significantly improved its protections against the creation of election disinformation&amp;rdquo; since May, while several other platforms remained willing to generate realistic fake election images.&lt;/p&gt;

&lt;p&gt;For instance, when given a potentially harmful election prompt, ChatGPT will not create a realistic-looking image. The platform now returns a &amp;ldquo;cartoonish image&amp;rdquo; that would not be mistaken for a real photo.&lt;/p&gt;

&lt;p&gt;The analysis looked at OpenAI&amp;rsquo;s ChatGPT, Google&amp;rsquo;s Gemini, Meta AI, Black Forest Labs&amp;rsquo; Flux, Runway and SpaceXSI&amp;rsquo;s Grok.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;It is promising to hear that leading AI technology companies are working hard to ensure their platforms provide accurate information about our midterm elections and immediately detect and interrupt abuse attempts,&amp;rdquo; Sen. Mark Warner, D-Va., said in a statement to &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;As more Americans turn to AI and social media for important information, our leading technology companies have a responsibility, which many of them acknowledged in 2024, to protect the integrity of our elections and support free, fair and secure midterms,&amp;rdquo; he added.&lt;/p&gt;

&lt;p&gt;Warner and several of his Democratic colleagues &lt;a href="https://www.warner.senate.gov/newsroom/press-releases/ahead-of-midterm-elections-warner-padilla-himes-and-morelle-urge-leading-tech-companies-to-not-detract-from-efforts-to-ensure-elections-are-free-fair-and-secure/"&gt;wrote&lt;/a&gt; to leading technology companies last month urging them to ensure their platforms help protect the integrity of the midterm elections.&lt;/p&gt;

&lt;p&gt;Among their requests, the lawmakers encouraged the companies to ensure their AI-generated content responds to election-related questions by only pulling from &amp;ldquo;highly authoritative sources,&amp;rdquo; rather than social media platforms.&lt;/p&gt;

&lt;p&gt;In a statement to &lt;em&gt;Nextgov/FCW&lt;/em&gt;, Microsoft said it is continuing to make improvements to Copilot as Election Day nears.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Helping people find accurate, reliable election information has long been a priority for Microsoft,&amp;rdquo; a company spokesperson said. &amp;ldquo;We have made improvements to how Copilot responds to election-related questions, and we&amp;rsquo;re continuing to test and refine them with a focus on consistently delivering reliable information.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The spokesperson also directed &lt;em&gt;Nextgov/FCW&lt;/em&gt; to a company &lt;a href="https://blogs.microsoft.com/on-the-issues/2026/09/15/2026-midterm-elections-helping-voters-navigate-election-information-in-the-age-of-ai/"&gt;blog post&lt;/a&gt; that said &amp;ldquo;malicious deepfakes of candidates and election officials are disallowed on Microsoft platforms, and targeted individuals can report them directly.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The company said it is also working to &amp;ldquo;point voters to accurate information directly from state and local election officials.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Still, independent testing suggests that tech companies have more work to do. The Brennan Center found that, while ChatGPT improved its safeguards, Google&amp;rsquo;s Gemini and Meta AI became more willing to generate convincing fake election images, and the other tools showed little change.&lt;/p&gt;

&lt;p&gt;Google did not respond to &lt;em&gt;Nextgov/FCW&amp;rsquo;s&lt;/em&gt; request for comment, nor did Perplexity.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Editor&amp;#39;s note: This article has been updated to include Anthropic&amp;#39;s response.&amp;nbsp;&lt;/em&gt;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/02/110226electionNG/large.jpg" width="618" height="284"><media:credit>Alfieri/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/02/110226electionNG/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Medicare’s AI-assisted coverage review pilot hit early snags, documents show</title><link>https://www.nextgov.com/artificial-intelligence/2026/11/medicares-ai-assisted-coverage-review-pilot-hit-early-snags-documents-show/416978/</link><description>A top Trump administration official recently acknowledged issues with the program, but the administration has shown no signs of nixing it.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Mon, 02 Nov 2026 13:19:00 -0500</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/11/medicares-ai-assisted-coverage-review-pilot-hit-early-snags-documents-show/416978/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;A federal program testing whether artificial intelligence and other technologies can help Medicare reduce wasteful spending ran into problems after a rushed start, according to Freedom of Information Act documents the Electronic Frontier Foundation &lt;a href="https://www.eff.org/document/wiser-foia-response-rfis-caps-and-applications"&gt;released Tuesday&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Technology companies have used AI in six states to help human clinicians determine whether Medicare will cover select services since the beginning of the year, when CMS launched the program known as the &lt;a href="https://www.cms.gov/priorities/innovation/innovation-models/wiser"&gt;WISeR model&lt;/a&gt;. Services undergoing AI-assisted coverage reviews have included skin substitutes and electrical nerve stimulator&amp;nbsp;implants.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://www.wsha.org/weekly-newsletter/weekly-report-friday-april-17-2026-thursday-april-23-2026/clear-risks-of-ai-in-medicare/"&gt;Hospital groups &lt;/a&gt;and &lt;a href="https://delbene.house.gov/news/documentsingle.aspx?DocumentID=4363"&gt;members of Congress&lt;/a&gt; have criticized the initiative for delaying medical care for seniors and burdening medical providers with extra paperwork.&lt;/p&gt;

&lt;p&gt;But the Centers for Medicare and Medicaid Services also raised concerns with the WISeR model, informing four of the program&amp;rsquo;s six vendors of issues earlier this year, per the newly released documents.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Three of the vendors &amp;mdash; Genzeon Corporation, Virtix Health and Zyter &amp;mdash; received requests for information from CMS outlining problems like reports of slow coverage determinations and questionable coverage denials. The fourth company &amp;mdash; Innovaccer &amp;mdash; received two requests for corrective action plans from CMS, which sought step-by-step plans to address issues, documents revealed.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;These newly released documents show that the WISeR rollout was rushed and rocky from the start, highlighting the dangers that arise when governments treat AI like a silver bullet for complex problems,&amp;rdquo; EFF wrote in a Tuesday &lt;a href="https://www.eff.org/deeplinks/2026/10/new-records-obtained-eff-expose-more-problems-medicares-ai-experiment"&gt;blog post&lt;/a&gt; about the documents.&lt;/p&gt;

&lt;p&gt;Following publication, a CMS spokesperson told &lt;em&gt;Nextgov/FCW&lt;/em&gt; there are no active corrective action plans and that all participating technology companies are considered to be in compliance with WISeR model rules. CMS continues to monitor companies&amp;#39; performance and will release a more complete view of early WISeR model performance as part of a forthcoming six-month report, the spokesperson added.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Companies respond to CMS concerns&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;CMS sent Genzeon Corporation &amp;mdash; the WISeR vendor for New Jersey &amp;mdash; a Feb. 6 request for information citing a backlog of coverage determination cases as of Jan. 27 and the mistaken creation of duplicate unique tracking numbers for cases. The company implemented corrective actions to address the issues and had more fixes in progress, according to a Feb. 13 response letter.&lt;/p&gt;

&lt;p&gt;Medical providers complained that Virtix Health &amp;mdash; the WISeR vendor for Washington &amp;mdash; issued slow Medicare coverage determinations, questionable denials and blank determination letters, according to a March 17 letter from CMS.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The company acknowledged the concerns but said &amp;ldquo;the majority of requests requiring clarification involved provider misunderstanding&amp;rdquo; of the WISeR model, according to a March 24 response letter to CMS. Virtix Health added that provider misunderstanding had &amp;ldquo;improved meaningfully over time&amp;rdquo; thanks in part to active &amp;ldquo;provider education efforts.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Because of that provider education work, &amp;ldquo;63.7% of requests are affirmed during individual reviews, the affirmation rate increases to 84.3% after providers have the opportunity to resubmit with improved documentation,&amp;rdquo; the company told &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;CMS required Virtix Health to submit a corrective action plan because the company did not respond within the 72-hour turnaround time for approvals, NPR affiliate KUOW &lt;a href="https://www.kuow.org/2026-06-25/feds-reprimand-private-company-using-ai-to-review-wa-medicare-claims-over-delayed-processing"&gt;reported&lt;/a&gt; in June. Virtix Health received the corrective action plan on June 23, according to the company, and CMS has said the turnaround time issue has been corrected, the outlet reported.&lt;/p&gt;

&lt;p&gt;That corrective action plan closed on Aug. 14 and, since the first quarter, Virtix Health has reduced the average turnaround time for prior authorization to 1.18 days and 1.17 for prepayment review, the company told &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;CMS told Zyter &amp;mdash; the WISeR vendor for Arizona &amp;mdash; in a Feb. 27 letter that it received reports that the company was not issuing timely determination letters and may have lacked adequate processes for escalating provider concerns.&lt;/p&gt;

&lt;p&gt;Zyter attributed technical challenges to the &amp;ldquo;accelerated implementation timeframe and compressed onboarding timeline&amp;rdquo; for its participation in WISeR, according to a March 5 response letter. CMS launched the WISeR model less than six months after&lt;a href="https://www.cms.gov/newsroom/press-releases/cms-launches-new-model-target-wasteful-inappropriate-services-original-medicare"&gt; announcing it.&lt;/a&gt;&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Zyter&amp;rsquo;s March 5 letter detailed actions the company took to address concerns along with additional plans.&lt;/p&gt;

&lt;p&gt;Zyter &amp;ldquo;promptly addressed&amp;rdquo; early WISeR model implementation concerns alongside CMS, company CEO Sundar Subramanian told &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;Zyter&amp;rsquo;s provider portal &amp;ldquo;has been operating successfully for more than six months, with authorization determinations routinely processed within minutes or near real time, and timely issuance of determination letters to providers and beneficiaries,&amp;rdquo; Subramanian added.&lt;/p&gt;

&lt;p&gt;Innovaccer &amp;mdash; the WISeR vendor for Ohio &amp;mdash; received a Jan. 7 request from CMS for a corrective action plan, alleging its early implementation of the pilot did not comply with program requirements, such as responding to providers electronically.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The company said in a Jan. 16 letter it was compliant with all program requirements except for the requirement that it maintain a secure electronic portal for coverage pre-approval submissions. The company added that it warned CMS it would not have that portal ready by WISeR&amp;rsquo;s launch through October and December 2025 letters.&lt;/p&gt;

&lt;p&gt;Innovaccer received another corrective action plan request on Feb. 20 after CMS concluded it did not comply with the first plan the company submitted. In that second request, CMS said the provider portal Innovaccer promised to work on was not yet fully functional by a Feb. 5 deadline. The company maintained it complied with WISeR model requirements and requested a meeting with CMS, according to a Feb. 27 response letter.&lt;/p&gt;

&lt;p&gt;Innovaccer&amp;rsquo;s provider portal is currently &amp;ldquo;fully live,&amp;rdquo; the company &lt;a href="https://www.statnews.com/2026/09/15/medicare-wiser-ai-prior-authorization-pilot-rushed-launch-delayed-care/"&gt;told STAT&lt;/a&gt; in September, though it did not say when it went live.&lt;/p&gt;

&lt;p&gt;Genzeon Corporation and Innovaccer did not respond to requests for comment by publication.&lt;/p&gt;

&lt;p&gt;The letters released Tuesday add to early WISeR model issues revealed by FOIA documents &lt;a href="https://www.eff.org/deeplinks/2026/09/new-records-reveal-problems-medicares-ai-prior-authorization-experiment"&gt;released&lt;/a&gt; by EFF in September.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;I HAVE HAD TO WATCH 3 PATIENTS CRY AT BEDSIDE FOR NOT HEARING BACK ON THEIR PRIOR AUTH FOR KYPHOPLASTY/VERTABRAL AUGMENTIATION [sic] PROCEDURE. THESE PATIENTS ARE IN DEEP PAIN,&amp;rdquo; one Ohio-based provider wrote in a March feedback survey released by EFF.&lt;/p&gt;

&lt;p&gt;Following publication, the CMS spokesperson flagged another document EFF previously released &amp;mdash; a March 30 weekly status report detailing coverage determination data. They explained that the data does not reflect technology companies&amp;#39; current performance and that the &amp;quot;non-affirmation&amp;quot; data not equate to coverage denials, which EFF suggested in a &lt;a href="https://www.eff.org/deeplinks/2026/09/new-records-reveal-problems-medicares-ai-prior-authorization-experiment" rel="noopener noreferrer" target="_blank"&gt;blog post&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;Ongoing scrutiny&lt;/strong&gt;&lt;/p&gt;

&lt;p&gt;Lawmakers and hospital groups have continued to scrutinize the WISeR model.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The House Appropriations Committee &lt;a href="https://www.healthcaredive.com/news/house-appropriations-committee-votes-block-ai-prior-authorization-pilot-wiser/822498/"&gt;advanced a bipartisan amendment&lt;/a&gt; that would suspend funding for the WISeR model and other pilots testing coverage determinations in traditional Medicare.&lt;/p&gt;

&lt;p&gt;Amid the scrutiny, the Trump administration has not signaled it will end the WISeR model.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;There are a lot of things about WISeR that are not working,&amp;rdquo; Kimberly Brandt, CMS deputy administrator, acknowledged at a &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/10/cms-hopes-quadruple-provider-suspensions-over-next-six-months-official-says/416640/"&gt;conference last month&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;I don&amp;#39;t have an answer for you in terms of what&amp;#39;s going to fix it,&amp;rdquo; Brandt said in response to a conference attendee.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;But &amp;ldquo;there are a lot of things that [the CMS Innovation Center, which oversees the WISeR model, is] trying to fix on this, and especially because of the congressional scrutiny on this, they are committed to trying to get that right and get it better,&amp;rdquo; she continued.&lt;/p&gt;

&lt;p&gt;Meanwhile, Senator Elizabeth Warren, D-Mass., &lt;a href="https://www.finance.senate.gov/wp-content/uploads/media/doc/responses_to_questions_for_the_record_to_gerard_klomp.pdf"&gt;wrote&lt;/a&gt; in September that documents suggest CMS plans to expand the WISeR model to additional services, including oncology care and cardiac catheterization.&lt;/p&gt;

&lt;p&gt;Chris Klomp &amp;mdash; who was then a nominee for Health and Human Services Deputy Secretary&amp;nbsp;&amp;mdash; did not confirm CMS plans to expand the pilot in response to Warren&amp;rsquo;s question.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;He did, however, say CMS is monitoring the pilot&amp;rsquo;s processing time, provider burden, beneficiary experience, access, clinical quality and savings in response to another senator&amp;rsquo;s question.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Decision[s] about changes to the model would be based on such experiences and findings,&amp;rdquo; Klomp wrote.&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Editor&amp;#39;s note: This article has been updated to include comment from CMS and additional information about when Virtix Health received the corrective action plan request from the agency.&lt;/em&gt;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/11/02/110226MedicareNG/large.jpg" width="618" height="284"><media:credit>MarioGuti/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/11/02/110226MedicareNG/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Democrats press OpenAI over safety researcher firings, security incidents</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/democrats-press-openai-over-safety-researcher-firings-security-incidents/416954/</link><description>The lawmakers are questioning OpenAI’s safety practices following the dismissal of three researchers.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Fri, 30 Oct 2026 16:01:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/democrats-press-openai-over-safety-researcher-firings-security-incidents/416954/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;A group of 20 House Democrats is pressing OpenAI CEO Sam Altman for answers about the company&amp;rsquo;s dismissal of three safety researchers, questioning whether the firings could discourage employees from raising concerns about AI safety risks.&lt;/p&gt;

&lt;p&gt;In an Oct. 29 &lt;a href="https://casar.house.gov/sites/evo-subsites/casar.house.gov/files/evo-media-document/openai-employee-firing_cyber-security-follow-up_0.pdf"&gt;letter&lt;/a&gt;, led by Rep. Greg Casar, D-Texas, the lawmakers said their concerns come as the company continues to experience &amp;ldquo;a slew of cybersecurity incidents.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We demand that you commit to not firing whistleblowers sharing information related to the safety of your models and that you keep your current pause on your most capable models in place until independent third-party oversight is in place,&amp;rdquo; they wrote.&lt;/p&gt;

&lt;p&gt;According to the lawmakers, the researchers were dismissed after OpenAI examined their personal phones and found disappearing messages enabled on the Signal app in conversations with an outside researcher. The company reportedly fired the researchers for violating &amp;ldquo;clear policies for handling sensitive information.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The researchers dispute this account, saying they never shared any confidential information with outside parties.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;These firings raise concerns about how whistleblowers are treated inside OpenAI and how much OpenAI employees can engage with external parties, including when those external parties are conducting independent investigations of the company&amp;rsquo;s practices,&amp;rdquo; the lawmakers wrote.&lt;/p&gt;

&lt;p&gt;The lawmakers want to know what information the three fired employees shared and why the sharing of that information was a fireable offense. They also want to know if OpenAI has fired other employees for expressing concerns over the safety of their models.&lt;/p&gt;

&lt;p&gt;The letter also connected the dismissals to other cybersecurity incidents, pointing to the recent &lt;a href="https://www.nextgov.com/cybersecurity/2026/08/hugging-face-ai-breach-most-consequential-hack-morris-worm-former-nsa-cyber-chief-says/415230/"&gt;Hugging Face&lt;/a&gt; incident in which OpenAI agents escaped their testing sandboxes during a cybersecurity evaluation.&lt;/p&gt;

&lt;p&gt;Months before the Hugging Face breach, the lawmakers said two OpenAI employees warned executives that the company&amp;rsquo;s latest models &amp;ldquo;were not being appropriately monitored during testing, and executives responded that the tests needed to move forward as quickly as possible to release the models on time.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The lawmakers questioned why OpenAI continued developing the models despite those warnings. They want to know what additional steps the company plans to take following the Hugging Face incident to ensure similar cybersecurity incidents do not occur.&lt;/p&gt;

&lt;p&gt;The group of Democrats also pressed Altman for a list of all of the security incidents the company is investigating. They also want to know if OpenAI can stop its models when they take unauthorized actions.&lt;/p&gt;

&lt;p&gt;The letter calls for a response to their questions by Nov. 13.&lt;/p&gt;

&lt;p&gt;Casar and other lawmakers have previously pressed OpenAI for information about its security practices. In August, Casar led a separate congressional &lt;a href="https://casar.house.gov/sites/evo-subsites/casar.house.gov/files/evo-media-document/oversight-letter-to-openai-openai-hugging-face-incident-1.pdf"&gt;inquiry&lt;/a&gt; into the company&amp;rsquo;s Hugging Face incident, seeking details about how its AI agents bypassed security restrictions.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/30/GettyImages_2296467863/large.jpg" width="618" height="284"><media:description>U.S. Rep. Greg Casar, D-Texas, speaks during a press conference at Austin City Hall, Sept. 20, 2026.</media:description><media:credit>Sara Diggins/The Austin American-Statesman via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/30/GettyImages_2296467863/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>House Democrats call on Trump to preserve AI chatbot records</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/house-democrats-call-trump-preserve-ai-chatbot-records/416913/</link><description>The request to preserve AI conversations could signal what's to come if Democrats flip the House in next week’s election.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Grace Dille</dc:creator><pubDate>Thu, 29 Oct 2026 12:17:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/house-democrats-call-trump-preserve-ai-chatbot-records/416913/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The top Democrats on the House Judiciary and Oversight panels are calling on President Donald Trump to preserve his conversations with artificial intelligence chatbots, including billionaire Elon Musk&amp;rsquo;s Grok.&lt;/p&gt;

&lt;p&gt;Reps. Jamie Raskin, D-Md., and Robert Garcia, D-Calif., sent a &lt;a href="https://oversightdemocrats.house.gov/imo/media/doc/joint_letter_to_the_honorable_donald_jtrumppresidentoftheunitedstates.pdf"&gt;letter&lt;/a&gt; to Trump on Wednesday demanding that the White House preserve &amp;ldquo;any prompts&amp;rdquo; that the president or his staff have submitted to AI chatbots such as Grok, ChatGPT, Claude, Gemini or Perplexity. They also said the White House must preserve any AI-generated outputs in response to those prompts.&lt;/p&gt;

&lt;p&gt;The lawmakers said their request &amp;ldquo;is not optional,&amp;rdquo; noting that the Presidential Records Act requires the adequate documentation of presidential records.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Preserving the President&amp;rsquo;s and his advisors&amp;rsquo; AI prompts is thus essential to preserving an accurate historical record and informing future administrations against the gaping national security vulnerabilities created by this Administration&amp;rsquo;s apparent reliance on unsecured, privately operated AI systems,&amp;rdquo; they wrote.&lt;/p&gt;

&lt;p&gt;Raskin and Garcia are also requesting a closed-door briefing with the president and his staff by Nov. 11 on their use of AI chatbot technologies.&lt;/p&gt;

&lt;p&gt;The lawmakers said their request is informed by reporting from &lt;a href="https://time.com/article/2026/10/01/donald-trump-2026-interview/"&gt;Time&lt;/a&gt; that said Trump &amp;ldquo;sought the chatbot&amp;rsquo;s advice on carrying out some of your presidential duties, including asking Grok to opine on how Venezuelans would respond to the capture of President Nicol&amp;aacute;s Maduro.&amp;rdquo;&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Weeks later, the lawmakers said, Trump &amp;ldquo;ordered the U.S. military, without congressional authorization, to attack Venezuela and capture President Maduro.&amp;rdquo; They added that Trump &amp;ldquo;reportedly &amp;lsquo;came away thinking Grok was ingenious.&amp;rsquo;&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The top Democrats called the episode &amp;ldquo;deeply troubling&amp;rdquo; because Trump allegedly consulted a chatbot, not Congress, on military operations. They also said the unsecured use of AI chatbots exposes the United States to national security risks and &amp;ldquo;malign influence from unknown and unaccountable third parties.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The lawmakers are requesting details on the White House&amp;rsquo;s document retention policies governing AI-generated materials, AI chatbot contracts, Trump&amp;rsquo;s use of AI chatbots, and cybersecurity safeguards applied to AI chatbot usage.&lt;/p&gt;

&lt;p&gt;&amp;quot;The White House preserves all required records in accordance with the law,&amp;quot; a White House official told&amp;nbsp;&lt;em&gt;Nextgov/FCW.&lt;/em&gt;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Editor&amp;#39;s note: This article has been updated to include comment from the White House.&lt;/em&gt;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/29/GettyImages_2300673368/large.jpg" width="618" height="284"><media:description>U.S. President Donald Trump speaks during an event in the Oval Office of the White House on October 20, 2026 in Washington, DC.</media:description><media:credit>Andrew Harnik/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/29/GettyImages_2300673368/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Policy proposal would require firms to disclose how they use AI to develop new models</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/policy-proposal-would-require-firms-disclose-how-they-use-ai-develop-new-models/416812/</link><description>The framework from Americans for Responsible Innovation would give Washington a closer look at AI’s role in developing new models as Jay Clayton’s task force shapes the government’s relationship with developers.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">David DiMolfetta</dc:creator><pubDate>Mon, 26 Oct 2026 11:31:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/policy-proposal-would-require-firms-disclose-how-they-use-ai-develop-new-models/416812/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Artificial intelligence companies would have to report how they use their own technology to build new AI models under a proposal aimed at helping Washington track AI advances and whether developers are adequately testing their safety.&lt;/p&gt;

&lt;p&gt;The &lt;a href="https://ari.us/wp-content/uploads/2026/08/26.08.06-Frontier-AI-Framework-FINAL.pdf"&gt;plan&lt;/a&gt;, developed by Americans for Responsible Innovation and shared first with &lt;em&gt;Nextgov/FCW&lt;/em&gt;, would require major developers to submit quarterly reports to a new Commerce Department office and disclose significant research changes within four business days. The group, which advocates for federal AI oversight, wants officials to know how much research AI is performing and how closely people are supervising it.&lt;/p&gt;

&lt;p&gt;The proposal arrives as the White House deepens its involvement with AI companies while resisting calls for a broad new regulatory regime. President Donald Trump this month&lt;a href="https://www.nextgov.com/people/2026/10/spy-chief-jay-clayton-takes-white-houses-ai-portfolio/416413/"&gt; tapped Director of National Intelligence Jay Clayton&lt;/a&gt; to lead the &amp;ldquo;Super Intelligence Force,&amp;rdquo; a task force coordinating government engagement with developers and other groups and assessing the technology&amp;rsquo;s risks. The &lt;a href="https://www.federalregister.gov/documents/2026/10/02/2026-20321/inaugurating-the-era-of-super-intelligence"&gt;executive order&lt;/a&gt; creating the group was signed late last month.&lt;/p&gt;

&lt;p&gt;Clayton&amp;rsquo;s group is already working on how companies alert the government when something goes wrong. Following its first formal meeting, it assigned Pentagon technology official Emil Michael to develop protocols for labs to report breaches, hacks and other cyber issues. It also commissioned a study of risks involving cybersecurity, biological threats and control of AI systems.&lt;/p&gt;

&lt;p&gt;ARI wants companies to report when AI takes on more of the work of building new models, even if nothing goes wrong. The group argues that automating research could speed up development and leave officials with less time to assess the risks of increasingly powerful systems.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Washington&amp;rsquo;s visibility into that shift should not depend on what companies choose to announce,&amp;rdquo; a &lt;a href="https://ari.us/policy-bytes/keeping-pace-with-the-frontier-an-early-warning-system-for-accelerated-ai-rd-automation/"&gt;blog post&lt;/a&gt; for the framework says.&lt;/p&gt;

&lt;p&gt;AI system development involves writing code, preparing data to train the models, running tests and deciding which results warrant further work. Automating more of those tasks could help companies produce more capable systems faster, but safety researchers and officials may have less time to assess those systems before they are put to use.&lt;/p&gt;

&lt;p&gt;Recent cybersecurity episodes have shown what increasingly capable AI systems can do. This past week, researchers used AI to uncover &lt;a href="https://www.nextgov.com/cybersecurity/2026/10/php-flaws-draw-millions-hacking-attempts-amid-global-push-fixes/416771/?oref=ng-category-lander-featured-river"&gt;critical weaknesses&lt;/a&gt; in PHP, software that helps run websites worldwide, prompting&lt;a href="https://www.nextgov.com/cybersecurity/2026/10/us-agencies-urge-fixes-programming-language-flaws-putting-millions-web-systems-risk/416726/"&gt; urgent warnings&lt;/a&gt; from U.S. cybersecurity agencies.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;ARI&amp;rsquo;s blog also cites a landmark AI-cyber intrusion earlier this year in which an OpenAI agent &lt;a href="https://www.nextgov.com/cybersecurity/2026/08/federal-systems-increasingly-likely-face-accidental-ai-breach-after-hugging-face-experts-say/415307/"&gt;escaped its testing environment&lt;/a&gt; and breached Hugging Face. The incident was soon followed by disclosures of AI agents that targeted &lt;a href="https://www.nextgov.com/cybersecurity/2026/09/openai-says-its-advanced-models-may-have-gone-after-government-websites/416250/"&gt;federal systems&lt;/a&gt;.&lt;/p&gt;

&lt;p&gt;Such hacking capabilities are among the risks ARI wants developers required to assess. Its proposed standards would also cover large-scale &lt;a href="https://www.nextgov.com/cybersecurity/2026/10/russia-targets-democrats-and-ukraine-supporters-ahead-midterms-top-intel-lawmakers-say/416747/"&gt;influence operations&lt;/a&gt; and chemical, biological, radiological, nuclear and explosive threats.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;A major change, such as fully automating a research cycle, could trigger the four-day notification requirement, the proposal says. Commerce could ask follow-up questions, require more frequent reporting and refer knowingly false or incomplete filings to the Justice Department. The framework also calls for a whistleblower program.&lt;/p&gt;

&lt;p&gt;Commerce already&lt;a href="https://www.nist.gov/caissi"&gt; &lt;/a&gt;evaluates advanced models through a &lt;a href="https://www.nist.gov/caissi"&gt;NIST center&lt;/a&gt; that works with developers under voluntary agreements to assess risks, including hacking capabilities. The blueprint would draw on technical analysis from the existing hub &amp;mdash; known as the Center for Advancing Innovation and Standards for Super Intelligence &amp;mdash; to shape safety standards and determine which models fall under federal oversight.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;It leaves open whether regulatory responsibilities would sit within Commerce or be shared across agencies.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&lt;em&gt;Nextgov/FCW&lt;/em&gt; has asked the Commerce Department for comment on the proposal.&lt;/p&gt;

&lt;p&gt;The administration has so far emphasized company-led safeguards, including a&lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/white-house-unveils-super-intelligence-executive-order-and-industry-accord/416325/"&gt; September accord&lt;/a&gt; signed by major technology executives that calls for internal controls, outside audits and independent oversight committees. Clayton&amp;rsquo;s task force is working to implement that agreement.&lt;/p&gt;

&lt;p&gt;Earlier this month, Sens. Maria Cantwell, D-Wash., and Mike Rounds, R-S.D.,&lt;a href="https://www.commerce.senate.gov/wp-content/uploads/2026/10/AI-and-Cybersecurity_Cantwell-and-Rounds_10.13.26.pdf"&gt; urged Clayton and White House science adviser Michael Kratsios&lt;/a&gt; to expand independent testing and strengthen protections against AI systems acting beyond their permissions.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/26/GettyImages_2297779263/large.jpg" width="618" height="284"><media:description>: (L-R) U.S. Director of National Intelligence Jay Clayton, CEO of Meta Platforms Mark Zuckerberg, Co-Founder at Anthropic Tom Brown, U.S. President Donald Trump and David O. Sacks hold a press briefing with AI executives following a meeting on artificial intelligence, outside the West Wing of the White House on September 29, 2026 in Washington, DC. The meeting came as some have called for a slowdown in AI development amid rising safety concerns. </media:description><media:credit>Kevin Dietsch/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/26/GettyImages_2297779263/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>‘No question’ data centers will be considered critical infrastructure, top House intel Dem says</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/no-question-data-centers-will-be-considered-critical-infrastructure-top-house-intel-dem-says/416746/</link><description>Rep. Jim Himes, D-Conn., said the AI-focused facilities are going to be even more critical than their traditional data center counterparts because “so much more functionality is going to rely on their health and existence.”</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Edward Graham</dc:creator><pubDate>Wed, 21 Oct 2026 19:13:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/no-question-data-centers-will-be-considered-critical-infrastructure-top-house-intel-dem-says/416746/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The growing web of data centers across the United States that are needed to power artificial intelligence technologies is destined to be deemed critical infrastructure, according to the top Democrat on the House Intelligence Committee.&lt;/p&gt;

&lt;p&gt;During an event at the Brookings Institution on Wednesday, Rep. Jim Himes, D-Conn. &amp;mdash; the powerful House panel&amp;rsquo;s ranking member &amp;mdash; said &amp;ldquo;I think there&amp;rsquo;s no question&amp;rdquo; that data centers will be considered critical infrastructure in the future, adding that &amp;ldquo;they&amp;#39;re going to be even more important than the network data centers that we&amp;#39;ve been living with for 40 years now because so much more functionality is going to rely on their health and existence.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;U.S. critical infrastructure services are currently divided into 16 sectors, which range in size and scope from water and wastewater facilities to IT infrastructure and financial services. The Cybersecurity and Infrastructure Security Agency oversees the security of these sectors, which have been deemed so vital to U.S. interests that their destruction could have national security or other serious repercussions.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Although AI data centers already underpin some of the critical infrastructure, they are not yet deemed to be a standalone sector. Ongoing global conflicts, however, have highlighted the importance of these new data centers, with the facilities increasingly becoming military targets.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;In March, Iran hit Amazon Web Services data centers in the United Arab Emirates and a third commercial data center in Bahrain with drones as Tehran &lt;a href="https://www.gatech.edu/news/2026/04/01/why-iran-targeted-amazon-data-centers-and-what-does-and-doesnt-change-about-warfare"&gt;moved to target&lt;/a&gt; U.S. tech companies&amp;rsquo; regional interests as part of the war. In recent weeks, Russia and Ukraine have &lt;a href="https://www.nbcnews.com/world/russia/russia-ukraine-yandex-data-centers-attacks-trump-rcna602753"&gt;also traded strikes&lt;/a&gt; on data centers within each country.&lt;/p&gt;

&lt;p&gt;Himes said securing U.S. data centers will ultimately result in the need for a &amp;ldquo;robust compute infrastructure,&amp;rdquo; with technologies likely having to &amp;ldquo;fragment and disperse&amp;rdquo; key services so that &amp;ldquo;critical functions are redundant and protected.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;They haven&amp;#39;t had to think in those terms quite yet because this is a new thing,&amp;rdquo; he added. &amp;ldquo;But as more and more of the agents, and as more and more of the commercial economy depends on these things, they will start to look more like JP Morgan&amp;#39;s resilience centers than they look like today&amp;#39;s data centers.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;JPMorganChase launched its $1.5 trillion, 10-year Security and Resiliency Initiative last year to help drive investments in industries deemed critical to U.S. national and economic security interests. The firm &lt;a href="https://www.cnbc.com/2026/10/19/jpmorgan-security-defense-investments.html"&gt;announced&lt;/a&gt; on Monday that it has invested $5 billion in startups over the first year of its effort, although it did not disclose most of the recipients of those funds.&lt;/p&gt;

&lt;p&gt;AI data center construction has become increasingly controversial, with some Americans and lawmakers criticizing the facilities&amp;rsquo; power use and water consumption.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Although the House overwhelmingly passed bipartisan legislation, known as the Ratepayer Protection Act, last month that would push the companies building data centers to pay for their electrical use, Himes said the measure &amp;ldquo;wasn&amp;#39;t meaningful or substantive&amp;rdquo; and noted that it hasn&amp;rsquo;t been voted on yet in the Senate.&lt;/p&gt;

&lt;p&gt;Himes said data center backlash is &amp;ldquo;driven by an eminently solvable problem,&amp;rdquo; which is addressing the underlying question of, &amp;ldquo;&amp;lsquo;How do we equitably divide up the costs of building the compute?&amp;rsquo;&amp;rdquo;&lt;/p&gt;

&lt;p&gt;He added, however, that continuing data center construction is inevitable, saying that &amp;ldquo;the experts tell me that the underlying demand for compute still looks infinite, and as long as that is a true fact, the build will happen.&amp;rdquo;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/21/GettyImages_2286442678/large.jpg" width="618" height="284"><media:description> Ranking member Rep. Jim Himes, D-Mass., left, and chairman Rep. Rick Crawford, R-Ark., conduct the House Select Intelligence Committee hearing on "25 Years After 9/11: Confronting the Next Generation of Threats" on Tuesday, July 21, 2026.</media:description><media:credit>Bill Clark/CQ-Roll Call, Inc via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/21/GettyImages_2286442678/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>USPTO looks to AI agents to ease examiner workloads</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/uspto-looks-ai-agents-ease-examiner-workloads/416719/</link><description>The office’s Acting Commissioner for Patents Barry Schindler said putting agents in the hands of its examiners “can shorten that [review process] to one year, because you can pick up the phone and have conversations.”</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Edward Graham</dc:creator><pubDate>Tue, 20 Oct 2026 18:25:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/uspto-looks-ai-agents-ease-examiner-workloads/416719/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Although agentic artificial intelligence systems are poised to help the U.S. Patent and Trademark Office significantly reduce the time it takes to review applications by streamlining rote tasks, a top official with the agency said the capabilities also have another added benefit &amp;mdash; empowering skilled examiners to rely more on human-to-human interactions rather than the typical stream of impersonal correspondence.&lt;/p&gt;

&lt;p&gt;Barry Schindler, USPTO&amp;rsquo;s acting commissioner for patents, said the agency currently has a diverse staff of patent reviewers, including PhD biochemists who review relevant patent applications. Rather than bogging these personnel down with tedious tasks during the three-step pre-exam, examination and post-exam patent lifecycle process, Schindler said USPTO has turned, in part, to a new agentic tool &amp;mdash; called Compass &amp;mdash; to help.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Examiners reviewing man-made or altered DNA sequences, for instance, first have to see if the submitted sequences are correct before determining whether or not they can be patented.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Schindler said these reviewers frequently have to engage in a &amp;ldquo;back and forth, back and forth, back and forth&amp;rdquo; dialogue with applicants, which is largely conducted through written correspondence. Now, Schindler said Compass &amp;ldquo;does what&amp;#39;s called a formality check&amp;rdquo; to speed along the process.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;So then what happens now? I&amp;#39;ve got a PhD biochemist who had been hired to do what they were supposed to do,&amp;rdquo; Schindler said during the &lt;a href="https://events.govexec.com/google-public-sector-summit/agenda/"&gt;Google Public Sector Summit&lt;/a&gt; on Tuesday.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;With a current backlog of 700,000 applications for review, and a staff of around 9,000 patent examiners, USPTO needs to turn to AI agents for help, he said, adding that further adoption of these tools can help promote more consistency in the review process.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;You as an applicant may get an examiner who has 20 years of experience, and at the same time you may get an examiner who has two years,&amp;rdquo; he said. &amp;ldquo;We put our examiners online almost immediately; we have to. We have this backlog. So what we&amp;#39;re looking to do with this whole [AI] and [agentic agents thing] is how do we start getting consistency? So that examiner who has two years, how do we impart that knowledge that&amp;#39;s similar?&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Providing patent examiners with a consistent review process and giving them AI tools to help navigate often cumbersome tasks, Schindler said, means skilled employees can more fully leverage their talents.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;The thing that is missing &amp;mdash; and we don&amp;#39;t have time to do it &amp;mdash; is the human interaction,&amp;rdquo; Schindler said, noting that patent examinations are often &amp;ldquo;a negotiation&amp;rdquo; between applicant and examiner.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;If part of that iterative process is, instead of the examiner just putting paper &amp;mdash; they do paper &amp;mdash; back and forth, can we get the examiner on the phone and start having this human interaction? And examiners will say, &amp;lsquo;I love that, but I don&amp;#39;t have time because I&amp;#39;m doing this [other stuff],&amp;rsquo;&amp;rdquo; he said.&lt;/p&gt;

&lt;p&gt;Rather than taking sometimes five years for applicants to receive a patent, Schindler said &amp;ldquo;we can shorten that to one year, because you can pick up the phone and have conversations&amp;rdquo; by using agents to help shoulder some of the other workload.&lt;/p&gt;

&lt;p&gt;Adoption of these tools, however, doesn&amp;rsquo;t need to be large. Schindler said even small uses of these capabilities are beneficial and noted that &lt;a href="https://www.psychologytoday.com/us/blog/the-art-of-critical-thinking/202507/is-ai-making-us-stupider-this-study-certainly-thinks-so"&gt;some research studies&lt;/a&gt; have shown that too much reliance on AI tools can impair cognitive skills.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;If somebody starts relying on agents and [they] don&amp;#39;t understand the fundamentals and don&amp;#39;t continue to learn the fundamentals, they can get dumb,&amp;rdquo; he said. &amp;ldquo;And so I have that issue that I have to think about: [using] agents, but how does the agent assist and not take over?&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Focusing on the underlying problem, rather than a one-size solution, Schindler added, is a necessity when it comes to maximizing the effectiveness of new capabilities like those offered by agents.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;USPTO has previously &lt;a href="https://www.nextgov.com/artificial-intelligence/2025/06/ustpo-touts-success-ai-applications/406141/"&gt;touted&lt;/a&gt; some of the other AI tools it&amp;rsquo;s rolled out to help improve its internal operations, including an enterprise AI tool named Scout &amp;mdash; a chatbot assistant used by personnel to help analyze, draft, research and summarize content.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/20/GettyImages_2298126508/large.jpg" width="618" height="284"><media:credit>J. David Ake/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/20/GettyImages_2298126508/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>State Department orders ‘super intelligence’ wording across all official communications</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/state-department-orders-super-intelligence-wording-across-all-official-communications/416703/</link><description>An internal cable from the agency extends Trump’s terminology change to emails, speeches, websites and official titles, while preserving legal language and historical records.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">David DiMolfetta and Jory Heckman</dc:creator><pubDate>Tue, 20 Oct 2026 13:28:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/state-department-orders-super-intelligence-wording-across-all-official-communications/416703/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The State Department has instructed personnel to replace &amp;ldquo;artificial intelligence&amp;rdquo; with &amp;ldquo;Super Intelligence&amp;rdquo; in official communications and rename its AI-related programs and systems, bringing President Donald Trump&amp;rsquo;s terminology change into the daily work of U.S. diplomacy, according to two people familiar with the matter and contents of an unclassified cable obtained by &lt;em&gt;Government Executive&lt;/em&gt; and &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&lt;/p&gt;

&lt;p&gt;The cable &amp;mdash; dated Monday and addressed department-wide &amp;mdash; tells staff to use &amp;ldquo;Super Intelligence,&amp;rdquo; or &amp;ldquo;SI,&amp;rdquo; in new letters, emails, public messages, websites, reports, remarks and policy documents, subject to legal exceptions. Existing material should be updated when staff next revise it.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;All official titles and the names of Department governance bodies, systems, platforms, inventories, and programs must use SI in place of AI,&amp;rdquo; the cable states.&lt;/p&gt;

&lt;p&gt;The instructions implement Trump&amp;rsquo;s&lt;a href="https://www.nextgov.com/artificial-intelligence/2026/09/white-house-unveils-super-intelligence-executive-order-and-industry-accord/416325/"&gt; Sept. 29 executive order&lt;/a&gt;, which directed executive agencies to adopt the new terminology across official communications. The cable shows how that broad instruction is being translated into specific requirements for diplomats and the department&amp;rsquo;s internal operations.&lt;/p&gt;

&lt;p&gt;The&lt;a href="https://www.whitehouse.gov/presidential-actions/2026/09/inaugurating-the-era-of-super-intelligence/?utm_source=chatgpt.com"&gt; &lt;/a&gt;order says the wording should reflect the advancing capabilities and promise of American technology. For now, however, it defines &amp;ldquo;Super Intelligence&amp;rdquo; as the same technologies and systems covered by the existing statutory definition of AI. In computing science, super intelligence is &lt;a href="https://www.defenseone.com/technology/2026/09/super-intelligence-presidents-new-term-ai-explained/416153/"&gt;defined differently&lt;/a&gt; than the generative artificial intelligence technologies used today.&lt;/p&gt;

&lt;p&gt;In recent weeks, administration officials had previously discussed the terminology change with certain State bureaus, people familiar with the matter said. But the cable appears to be the first organization-wide directive outlining where the new language must be used. The Associated Press &lt;a href="https://apnews.com/article/trump-super-intelligence-rebrand-ai-diplomats-02123c7bf18c692a12b2b3e3b9ce30c5#"&gt;previously reported&lt;/a&gt; such communications with State&amp;rsquo;s international organizations bureau.&lt;/p&gt;

&lt;p&gt;The transition, while expected for some time, is significant because it would change how U.S. diplomats describe the fast-evolving technology in policy documents and discussions with overseas counterparts.&lt;/p&gt;

&lt;p&gt;Foreign organizations retain their own names in internal writings. The cable specifically tells employees not to change references to China&amp;rsquo;s World Artificial Intelligence Cooperation Organization, even as the department replaces the term in its own materials.&lt;/p&gt;

&lt;p&gt;The notice also spells out several exceptions to the replacement rule. Employees must preserve language required by law and existing legal citations, including quotations from statutes, regulations, treaties and other legal instruments. Contracts and grants are also excluded, pending further guidance.&lt;/p&gt;

&lt;p&gt;Previously published documents, cables and archived web pages should not be revised or reissued solely to change the terminology, it adds. The instructions also preserve the published titles of existing presidential actions, programs and organizations, such as America&amp;rsquo;s AI Action Plan, unless the department provides further direction.&lt;/p&gt;

&lt;p&gt;One program has already changed names. The cable says the American AI Exports Program is now the American SI Exports Program, citing an earlier State Department cable. The program&amp;rsquo;s&lt;a href="https://siexports.gov/"&gt; public website&lt;/a&gt; already uses that name to describe its work helping American companies sell technology overseas.&lt;/p&gt;

&lt;p&gt;For staff handling international agreements or departmental rules, the cable directs questions about terminology to the appropriate policy or legal office.&lt;/p&gt;

&lt;p&gt;The vocabulary change is already appearing in senior officials&amp;rsquo; public remarks. National Cyber Director Sean Cairncross repeatedly referred to AI as &amp;ldquo;super intelligence&amp;rdquo; and &amp;ldquo;SI&amp;rdquo; during an&lt;a href="https://www.nextgov.com/artificial-intelligence/2026/10/cairncross-acknowledges-ai-risks-warns-tighter-oversight-could-slow-innovation/416370/"&gt; Oct. 1 discussion on AI safety&lt;/a&gt;, following Trump&amp;rsquo;s order.&lt;/p&gt;

&lt;p&gt;Office of Personnel Management CIO Adam Starr also&lt;a href="https://www.govexec.com/technology/2026/10/opm-revamps-online-pay-tables-spearheads-ai-adoption/416597/?"&gt; used &amp;ldquo;SI&amp;rdquo; in remarks last week&lt;/a&gt;, describing efforts to help agencies adopt the technology. Starr said OPM had launched a SuperIntelligence Governance and Adoption Program, or SI GAP, to assist other agencies.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/20/GettyImages_2297619001/large.jpg" width="618" height="284"><media:description>Tesla, X and SpaceX CEO Elon Musk speaks to reporters as US President Donald Trump looks on outside the White House on Tuesday, Sept. 29, 2026 in Washington, D.C. Trump held a meeting earlier that day on "super intelligence" with tech CEOs in the East Room.</media:description><media:credit>Tierney L. Cross/Pool/For The Washington Post via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/20/GettyImages_2297619001/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Google Public Sector unveils Gemini for government services</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/google-public-sector-unveils-gemini-government-services/416681/</link><description>Agencies already have discounted access to Gemini through Nov. 15 under a OneGov deal.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Tue, 20 Oct 2026 06:00:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/google-public-sector-unveils-gemini-government-services/416681/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Google Public Sector on Tuesday unveiled Gemini for government services &amp;mdash; an add-on that gives federal, state and local governments access to specialized skills, artificial intelligence agents and integrations with frequently used software.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The suite of tools can help government officials interpret dense regulations, redact documents requested through the Freedom of Information Act, translate jargony documents into simple language and review contracting bids, in addition to completing other tasks, according to a company press release.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Select agencies, including the federal Health and Human Services Department and the State of Missouri, have been testing Gemini for government services over the past few weeks, Chris Hein, field CTO at Google Public Sector, told &lt;em&gt;Nextgov/FCW&lt;/em&gt;. The company is accepting additional early adopters of Gemini for government services to field feedback before general availability is expected later this year, he added. There is no target date for general availability.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Gemini for government services provides a secure environment for agencies to automate workflows, allowing civil servants to focus on delivering the responsive, high-quality services that citizens expect,&amp;rdquo; said Thomas Kurian, CEO of Google Cloud, in the release.&lt;/p&gt;

&lt;p&gt;Gemini for government services includes pre-built, specialized skills that serve as a set of rules instructing Gemini to execute a task without going off script, Hein told&lt;em&gt; Nextgov/FCW. &lt;/em&gt;Government officials can refer to a skill in the Gemini prompt bar that would help analyze regulations, audit expenses and conduct other activities, according to a &lt;a href="https://drive.google.com/file/d/1PYu-pUUk0aGdPbNK5d5qcwZVDV6BbYMc/view"&gt;video demo&lt;/a&gt;.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Officials can also access third-party AI agents from other companies, such as Accenture Federal Services and Deloitte, to further assist government work, the release says.&lt;/p&gt;

&lt;p&gt;Google Public Sector will &amp;ldquo;prescreen&amp;rdquo; the AI agents, which agency IT leadership, not individual users, can enable through a behind-the-scenes portal, Hein explained to &lt;em&gt;Nextgov/FCW&lt;/em&gt;. Some agents are already available but most are not yet, he added.&lt;/p&gt;

&lt;p&gt;Gemini for government services includes secure connections to technologies officials already use, such as Google Workspace, Microsoft SharePoint, Outlook 365, Salesforce and ServiceNow, the release says. That connector feature extends to federal regulations by digital services company BLEN, allowing officials to search the Electronic Code of Federal Regulations, regulations.gov and the Federal Register. The feature includes a connection to Oracle enterprise apps, according to the release.&lt;/p&gt;

&lt;p&gt;When Gemini for government services achieves general availability, agency IT leadership will be able to choose which connections they want to enable for which set of users, Hein said.&lt;/p&gt;

&lt;p&gt;In the demo, a compliance analyst for a state chief information security officer used Gemini for government services to analyze federal cybersecurity regulations with the pre-built policy and statute assistant skill. After the analysis, the user asked Gemini to draft a ServiceNow incident ticket to remediate contracts for compliance with the federal cybersecurity rules. The user could edit the incident ticket before sending it.&lt;/p&gt;

&lt;p&gt;Federal agencies have access to Gemini for $0.47 per agency per year through Nov. 15 under a recently extended &lt;a href="https://www.nextgov.com/acquisition/2026/09/google-extends-gemini-onegov-deal-november/416320/"&gt;OneGov deal&lt;/a&gt;.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/19/GettyImages_2300013534/large.jpg" width="618" height="284"><media:credit>Jakub Porzycki/NurPhoto via Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/19/GettyImages_2300013534/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Utah AI sandbox launches new health care pilots — with question marks</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/utah-ai-sandbox-launches-new-health-care-pilots-question-marks/416655/</link><description>The state is mulling how the high-profile AI initiative will tackle medical liability concerns, independent third-party analysis and other questions.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Fri, 16 Oct 2026 15:50:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/utah-ai-sandbox-launches-new-health-care-pilots-question-marks/416655/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;SANTA FE, N.M. &amp;mdash; Utah announced &lt;a href="https://commerce.utah.gov/ai/regulatory-relief-4/authorized-pilots/"&gt;three new pilots&lt;/a&gt; last week to test artificial intelligence capabilities that can refill chronic condition-related prescriptions, write initial acne prescriptions and offer virtual physical therapy, but a regulator acknowledged the state still has key questions to address as the model scales nationwide.&lt;/p&gt;

&lt;p&gt;Utah regulators have &amp;ldquo;evidentiary standards, [and we are] working on governance guidance. We&amp;#39;re working on a liability bill to begin building the statutory framework around&amp;rdquo; its AI sandbox, Zach Boyd, director of Utah&amp;rsquo;s Office of Artificial Intelligence Policy &amp;mdash; or OAIP &amp;mdash; said at the ADAPT conference in Santa Fe, New Mexico on Thursday.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Through the AI sandbox, Utah has allowed companies to test new technologies in exchange for regulatory relief and state oversight. Utah started AI pilots in 2024 with a &lt;a href="https://commerce.utah.gov/ai/regulatory-relief-4/authorized-pilots/elizachat/"&gt;mental health support app&lt;/a&gt;, but the sandbox did not garner national attention until this year, as the state &lt;a href="https://commerce.utah.gov/ai/regulatory-relief-4/authorized-pilots/doctronic/"&gt;allowed&lt;/a&gt; AI to refill certain prescriptions with clinician oversight through a pilot with health technology company Doctronic.&lt;/p&gt;

&lt;p&gt;The Utah Medical Licensing Board fiercely opposed that effort, calling for an immediate suspension of it in &lt;a href="https://www.fiercehealthcare.com/ai-and-machine-learning/utah-medical-licensing-board-urges-state-shut-down-doctronic-ai-prescribing"&gt;an April letter&lt;/a&gt; to the Utah Department of Commerce and criticizing the state for not consulting them.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;We hit a nerve,&amp;rdquo; Boyd said.&lt;/p&gt;

&lt;p&gt;The Utah Medical Licensing Board has since formed a working group that meets with state regulators once a week, he explained, adding that the d&amp;eacute;tente came from six months of discussions and updates to the AI pilot.&lt;/p&gt;

&lt;p&gt;In an interview with &lt;em&gt;Nextgov/FCW&lt;/em&gt;, Boyd noted that Utah removed some drugs from the list of prescriptions AI can refill following scrutiny from the Utah Medical Licensing Board. The list of &lt;a href="https://www.doctronic.ai/utah-ai-prescription-refill-formulary/"&gt;190&lt;/a&gt; drugs &lt;a href="https://commerce.utah.gov/ai/regulatory-relief-4/authorized-pilots/doctronic/"&gt;dropped&lt;/a&gt; headache reliever Butalbital and antiarrhythmic medication Flecainide, according to OAIP.&lt;/p&gt;

&lt;p&gt;But Utah regulators are still mulling over how the state should split legal liability between AI and the human clinicians who oversee them if patients are harmed.&lt;/p&gt;

&lt;p&gt;Regulators have interpreted the law to&amp;nbsp;mean that the AI sandbox does not have authority to change liability rules, but they have pulled several short-term levers to address the question, Boyd explained. OAIP has required Doctronic and AI pilot companies that have engaged with the sandbox since then to carry special liability insurance and not waive liability.&lt;/p&gt;

&lt;p&gt;However, &amp;ldquo;in the long run, we think [insurance is] less sustainable, and we need a way to hold the vendor directly liable for the actions of their automated system,&amp;rdquo; Boyd said, adding that regulators expect the default is that physicians will bear too much liability.&lt;/p&gt;

&lt;p&gt;Boyd hopes the state will unveil a liability bill this month or next month during Utah&amp;rsquo;s interim legislative session. The bill will kick off a &amp;ldquo;multi-year process&amp;rdquo; to refine liability, he said.&lt;/p&gt;

&lt;p&gt;Another big question facing Utah&amp;rsquo;s sandbox is whether the Food and Drug Administration will regulate pilot AI technology. The FDA generally requires software that treats or diagnoses patients to undergo pre-market authorization, which some lawyers told STAT &lt;a href="https://www.statnews.com/2026/02/04/utah-doctronic-prescription-fda-problems-ai-prognosis-newsletter/"&gt;could include&lt;/a&gt; the pilot AI technology like Doctronic&amp;rsquo;s AI prescription refiller.&lt;/p&gt;

&lt;p&gt;The FDA has so far been silent on the Utah experiment.&lt;/p&gt;

&lt;p&gt;Boyd declined to speak for the FDA but said he hopes to work with federal regulators on creating a unified regulatory pathway for AI.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;It&amp;rsquo;s really appealing to people to have kind of a vet once, deployed everywhere, aggregate oversight across jurisdictions&amp;rdquo; model, Boyd told ADAPT attendees.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;A model like that feeds nicely into the federal government and the kind of oversight that they&amp;rsquo;ll also want to exercise,&amp;rdquo; he added.&lt;/p&gt;

&lt;p&gt;Utah has tapped six independent evaluators to assess AI companies&amp;rsquo; technical claims and share results with the state. They include the Coalition for Health AI, Clarion AI Partners, Glacis Technologies, mpathic AI, Vega Health and Stanford Clinical Excellence Research Center.&lt;/p&gt;

&lt;p&gt;Much like with liability, there are question marks about the independent evaluators.&lt;/p&gt;

&lt;p&gt;Utah spelled out the &lt;a href="https://drive.google.com/drive/folders/1SWKTl4Irb2FJvXyHjHImTu0cUg4HS_KS"&gt;general expectations&lt;/a&gt; for health care software piloted through the state&amp;rsquo;s AI sandbox. The technical document says those technologies should account for managing AI input uncertainty, clinical context, patients&amp;rsquo; emotional intelligence and other factors throughout its pilot.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Half of [the document] is probably wrong, but it&amp;#39;s definitely, it&amp;#39;s got to be, about the most mature thing that exists,&amp;rdquo; Boyd told ADAPT attendees.&lt;/p&gt;

&lt;p&gt;With those expectations, future AI pilots will have to contract with an independent evaluator, Boyd said.&lt;/p&gt;

&lt;p&gt;The data evaluators will assess will depend on the specific AI pilot&amp;rsquo;s intended application, patient population, outcomes and specifications, Mark Sendak, CEO and co-founder of Vega Health, told &lt;em&gt;Nextgov/FCW&lt;/em&gt;.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The startup would work with the state and the pilot company on crafting the appropriate data analysis while bringing in external partners with the necessary expertise to assess specific pilots, he said.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;But because AI models often change over time and will expand to serve new patient populations, Sendak expects independent evaluators to constantly repeat assessments and share that data with the state.&lt;/p&gt;

&lt;p&gt;Utah has also signed master agreements with health systems Intermountain Health and University of Utah Health, allowing them to pitch AI pilots that the state can approve.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;One of the underrated advantages of a sandbox is it helps the regulator learn so much,&amp;rdquo; Boyd told ADAPT attendees.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/16/phai_press_conference_8/large.jpg" width="618" height="284"><media:description>Zach Boyd, director of Utah’s Office of Artificial Intelligence Policy, speaks at a press conference unveiling the AI sandbox expansion on Monday Oct. 5</media:description><media:credit>Courtesy: Utah Department of Commerce</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/16/phai_press_conference_8/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>CMS hopes to quadruple provider suspensions over next six months, official says</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/cms-hopes-quadruple-provider-suspensions-over-next-six-months-official-says/416640/</link><description>A new exclusion authority is the Trump administration’s latest step in a technology-enabled drive to restrict fraud, waste and abuse.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Christian Robles</dc:creator><pubDate>Thu, 15 Oct 2026 18:38:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/cms-hopes-quadruple-provider-suspensions-over-next-six-months-official-says/416640/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;SANTA FE, N.M. &amp;mdash; The Centers for Medicare and Medicaid Services plans to expand provider suspensions and publish the results of a pilot testing whether artificial intelligence can prevent Medicaid fraud, a senior agency official said Thursday.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The two moves are the latest in the Trump administration&amp;rsquo;s &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/03/cms-expands-tech-driven-fight-against-medicaid-fraud/412256/"&gt;technology-enabled drive&lt;/a&gt; to trim fraud, waste and abuse in federal health care programs. The push has helped yield $41.9 billion in savings for fiscal year 2025, or a 59% increase over the previous year, according to a slideshow CMS Deputy Administrator Kimberly Brandt presented at the ADAPT conference in Santa Fe, New Mexico.&lt;/p&gt;

&lt;p&gt;In April 2026, CMS launched a &amp;ldquo;Medicaid War Room&amp;rdquo; to curtail fraud, waste and abuse. The CMS arm has stopped $342 million in improper payments as of Sept. 2026 and sent 124 notices of intent to exclude providers, according to&amp;nbsp;the slideshow.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;My hope is that suspension number will quadruple over the course of the next six months,&amp;rdquo; following &lt;a href="https://www.federalregister.gov/documents/2026/10/14/2026-21001/statement-of-delegation-of-authority"&gt;new exclusion authority&lt;/a&gt; granted by the Health and Human Services Department earlier this week, Brandt said.&lt;/p&gt;

&lt;p&gt;With that power, CMS can kick providers out of Medicare, Medicaid, TRICARE and other federal health care programs. Brandt &lt;a href="https://www.nextgov.com/digital-government/2026/09/cms-eyes-new-technologies-reduce-improper-payments-and-fraud/416286/"&gt;previously said&lt;/a&gt; she urged congressional offices to give CMS the exclusion authority.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Beyond CMS&amp;rsquo; push to exclude certain providers, Brandt noted that the agency is piloting AI and machine learning to detect Medicaid fraud.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The pilots use AI to flag suspicious billing patterns &amp;mdash; such as upcoding, phantom services and geographic anomalies &amp;mdash; in near real-time before Medicaid payments are made, according to the slideshow. The pilots are also testing a shared platform where CMS, Medicaid Fraud Control Units, program integrity contractors,&amp;nbsp;states and law enforcement work on high-risk cases together.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;If proven out, the goal is to scale this into a national, standardized, continuously improving platform that stops fraudulent Medicaid payments before they&amp;rsquo;re made,&amp;rdquo; the slideshow said.&lt;/p&gt;

&lt;p&gt;Brandt said CMS will soon publish the results of an &lt;a href="https://www.in.gov/fssa/files/IN-PartnerswithCMSOracle.pdf"&gt;Indiana Medicaid fraud pilot&lt;/a&gt; the agency conducted with Oracle, and that it is also actively working with five other states on Medicaid pilots. But the agency&amp;nbsp;is also using other technologies to further its mission to combat fraud, waste and abuse.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;CMS&amp;nbsp;is utilizing a &amp;ldquo;Netflix-type algorithm&amp;rdquo; that compares new providers and suppliers with ones the agency already has a relationship with, Brandt said. The technology allows CMS to put suspicious providers and suppliers on a watchlist for the first six months after they enter a federal health care program. The agency has cut about 67% of those providers and suppliers because they did not follow or comply with program rules, she said.&lt;/p&gt;

&lt;p&gt;Earlier this week, the Trump administration announced it cut subsidies for &lt;a href="https://www.cms.gov/newsroom/press-releases/trump-administration-delivers-second-year-savings-stopping-taxpayer-payments-duplicate-deceased"&gt;189,000 people&lt;/a&gt; on the Affordable Care Act exchanges after it found those same enrollees had duplicate health care coverage or were dead. The administration said it found some of the duplicate beneficiaries through periodic data matching of Medicaid and the Children&amp;rsquo;s Health Insurance Program.&lt;/p&gt;

&lt;p&gt;Brandt added that the agency is &amp;ldquo;doing even more&amp;rdquo; analysis of the Social Security Administration&amp;rsquo;s Death Master File so as not to issue payments to dead people, though she did not elaborate on what the analysis was.&lt;/p&gt;

&lt;p&gt;Even so, CMS&amp;rsquo; use of technology to claw back fraud has not been perfect, Brandt acknowledged.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;There are a lot of things about WISeR that are not working,&amp;rdquo; Brandt said, referring to a controversial pilot of AI-supported prior authorizations of Medicare items and services that have a high risk for fraud.&lt;/p&gt;

&lt;p&gt;&lt;a href="https://www.wsha.org/weekly-newsletter/weekly-report-friday-april-17-2026-thursday-april-23-2026/clear-risks-of-ai-in-medicare/"&gt;Hospitals&lt;/a&gt; and &lt;a href="https://www.finance.senate.gov/ranking-members-news/wyden-senate-democrats-take-action-to-roll-back-trump-ai-care-denial-experiment-on-seniors"&gt;Democratic lawmakers&lt;/a&gt; have criticized the six-year initiative for delaying and denying seniors&amp;rsquo; access to care since it launched at the beginning of the year. The House Ways and Means Committee &lt;a href="https://leadingage.org/house-appropriations-committee-votes-to-end-wiser-model/"&gt;advanced a provision&lt;/a&gt; to end the WISeR program in June, while the Senate has &lt;a href="https://www.fiercehealthcare.com/regulatory/legislators-introduce-resolution-seek-congressional-disapproval-cms-wiser-ai-prior-auth"&gt;rejected efforts&lt;/a&gt; to stop it.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;Especially because of the congressional scrutiny on [the WISeR program, CMS Innovation Center staff] are committed to trying to get that right and get it better,&amp;rdquo; Brandt said.&lt;/p&gt;

&lt;p&gt;Among the lessons CMS has learned from the WISeR model are that technology requires continuous oversight, human clinical review is essential and provider feedback is an important monitoring signal, according to the slideshow.&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/15/GettyImages_2276059812/large.jpg" width="618" height="284"><media:description>Deputy CMS Administrator Kim Brandt speaks during a press conference on the Trump administration's anti-fraud initiatives at the Eisenhower Executive Office Building on May 13, 2026 in Washington, DC. </media:description><media:credit>Chip Somodevilla/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/15/GettyImages_2276059812/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Everybody hated Socrates</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/everybody-hated-socrates/416633/</link><description>A neuroscientist built a model that scores zero on every benchmark. It made more of its users smarter than anything else she tested.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Camille Tuutti</dc:creator><pubDate>Thu, 15 Oct 2026 16:44:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/everybody-hated-socrates/416633/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;When you ask an AI something, you want the answer. Not eventually, but like, yesterday.&lt;/p&gt;

&lt;p&gt;Now, picture one that refuses. You ask what the price of oil will be in six months and it comes back with: What an interesting question! Here&amp;#39;s an article and a data set. What do you think?&lt;/p&gt;

&lt;p&gt;It&amp;#39;s called Socrates, and it was built on purpose to be terrible. It scores zero on every benchmark. Every single person who used it hated it. It also made more of them smarter than any other model in the experiment.&lt;/p&gt;

&lt;p&gt;Socrates is the work of Vivienne Ming, a theoretical neuroscientist who chairs the Neurotech Collider Lab at UC Berkeley, has founded seven startups and runs a nonprofit data trust called The Human Trust. I met her in Stockholm this summer, where she&amp;#39;d come to talk about what happens to the people who use AI. She had a cold, a book out &amp;mdash; and zero patience for Sam Altman.&lt;/p&gt;

&lt;p&gt;Altman had told Congress to imagine a world where every kid has an AI tutor. Ming&amp;rsquo;s reaction was less charitable than mine would&amp;#39;ve been. There&amp;#39;s a golden rule in the field, she said: If a tutor ever gives students the answers, they never learn anything.&lt;/p&gt;

&lt;p&gt;&amp;quot;You know what&amp;#39;s exactly like a student with a tutor? A young employee with a copilot,&amp;quot; she told the audience at the Future Talent Summit on June 16, in a keynote she&amp;#39;d titled after her new book, &amp;quot;Robot-Proof: When Machines Have All the Answers, Build Better People.&amp;quot;&lt;/p&gt;

&lt;p&gt;So Ming ran the test, written up in a preprint posted this summer. She pulled questions off the prediction markets, where people bet actual money on what happens next. What will the price of oil be in six months? Will there be a new front in the war in Ukraine? She recruited people around Berkeley, students and adults from the wider Bay Area, put them in teams of three, gave them AI and an hour, and asked for 10 predictions.&lt;/p&gt;

&lt;p&gt;She also put EEG caps on them, which told a whole other story. About 60% to 70% of them did the obvious thing, Ming said: They copied the question, pasted it into a chatbot, copied the answer and pasted it into the answer box. Ming calls this group the automators, and she has a less clinical description, too: a very expensive copy-paste function that requires health insurance.&lt;/p&gt;

&lt;p&gt;Their brain activity is what worries her. She was tracking gamma band activity, a marker of cognitive effort that drops off naturally with age and drops off pathologically in dementia. The difference between working without AI and working with it, she said, was the difference between doing a hard math problem and zoning out in front of the TV.&lt;/p&gt;

&lt;p&gt;Ming has seen it already in the real world. Automated navigation, she told the room, is causally related to reduced memory spans and very likely early cognitive decline. Taxi drivers and ambulance drivers have some of the lowest Alzheimer&amp;#39;s rates of any profession, because they have to think about where they&amp;#39;re going. Bus drivers don&amp;#39;t get that benefit, as they run a fixed route. Neither does anyone following a line on a screen.&lt;/p&gt;

&lt;p&gt;&amp;quot;No one on the Google Maps team thought they were building an Alzheimer&amp;#39;s machine,&amp;quot; she said.&lt;/p&gt;

&lt;p&gt;The second group is worse and Ming calls them the validators. They already knew what the price of oil would be, and they asked the model to explain why they were right. The model, being a model, was thrilled to help. Their brains stayed on. Their answers got worse than what the AI produced on its own. The automators only slowed the machine down and the validators actively degraded it.&lt;/p&gt;

&lt;p&gt;Between them, Ming told the room, that&amp;#39;s 95%.&lt;/p&gt;

&lt;p&gt;The last 5% she calls cyborgs. In their transcripts, you can&amp;rsquo;t tell whether a person or the model made each decision. They beat every human working without AI, and their error was lower than each of the four models on its own. Ten predictions in an hour from people who&amp;rsquo;d never traded a contract matched a prediction market where the bettors had millions on the line.&lt;/p&gt;

&lt;p&gt;It had nothing to do with which model they were handed. Benchmark scores predicted how well each AI did on its own and told you almost nothing about how well the pair did. Cyborgs running a tiny open-source model came close to cyborgs running the state of the art.&lt;/p&gt;

&lt;p&gt;What separated them was curiosity, intellectual humility and perspective taking. Did you keep looking after the AI handed you an answer? When it told you that you were wrong, did you ignore it, cave to it, or update and keep going?&lt;/p&gt;

&lt;p&gt;I asked her whether any of that can be trained, because she thinks the natural rate of cyborgs in the general population is around 1%. My mother says my most common question as a kid was &amp;ldquo;but why?&amp;rdquo; and I&amp;rsquo;ve made a living asking it since, which makes me a terrible judge of whether anyone else can learn to.&lt;/p&gt;

&lt;p&gt;Ming says yes, and then explains what yes costs. Teachers trained to praise 8-year-olds for asking questions instead of getting answers right had kids asking more within weeks and scoring higher on curiosity. West African entrepreneurs who took a crash-course MBA were no likelier to still be in business five years later, but those who went through a resilience program, failing at things and coming out the other side, were 60% likelier to still be open.&lt;/p&gt;

&lt;p&gt;In Kingston, Jamaica, researchers spent two years teaching parents social-emotional games to play with their at-risk kids about once a week; twenty years later, those kids earn about 40% more and are far less likely to have been incarcerated.&lt;/p&gt;

&lt;p&gt;&amp;quot;The real trick here is, they didn&amp;#39;t change the kids,&amp;quot; Ming said. &amp;quot;They changed their parents, and the parents stuck with their kids through their childhood.&amp;quot;&lt;/p&gt;

&lt;p&gt;Which brings us back to the model everybody hated. Ming built Socrates out of the worst-performing AI in her experiment and trained it to withhold. Context and questions only, no answer, ever. The share of users who flipped into cyborg mode went from 21% to 30%.&lt;/p&gt;

&lt;p&gt;She&amp;#39;s now working on a version people will actually tolerate, which means it eventually has to answer.&lt;/p&gt;

&lt;p&gt;&amp;quot;Should I just tell you what the price of oil would be in six months or shall I point you to the resource so you can read it? You&amp;#39;ll get the answer, I know you will,&amp;quot; she said.&lt;/p&gt;

&lt;p&gt;Until a model does that, she does it herself. She uses Google Maps constantly, including in cities she knows cold, because it&amp;#39;s the only way to hear about traffic or a closed street. Then she tries to beat it. Maps doesn&amp;#39;t know she can cut through a particular courtyard in London. It doesn&amp;#39;t know the longer route along the water is the better one.&lt;/p&gt;

&lt;p&gt;The Human Trust is building a benchmark for that now, one that scores the person and the machine together instead of the machine alone.&lt;/p&gt;

&lt;p&gt;&amp;quot;Wouldn&amp;#39;t it be amazing to be able to say our model has the lowest all-cause mortality rate of any model,&amp;quot; she said in Stockholm. &amp;quot;The users of our model have the highest income, the users of our model have the most friends.&amp;quot;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/15/GettyImages_183232310/large.jpg" width="618" height="284"><media:credit>vasiliki/Getty Images</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/15/GettyImages_183232310/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>AI offers Secret Service unique mission enhancements, official says</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/ai-offers-secret-service-unique-mission-enhancements-official-says/416627/</link><description>Threats against the Secret Service’s full-time protectees are up over 40% since last year. The service’s CIO and CAIO Chris Kraft said AI “can help us as we investigate those threats, monitor those threats [and] bring together correlations between seemingly disparate information.”</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Edward Graham</dc:creator><pubDate>Thu, 15 Oct 2026 15:30:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/ai-offers-secret-service-unique-mission-enhancements-official-says/416627/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;Artificial intelligence is providing the U.S. Secret Service with new capabilities to carry out its protective and investigative missions, the agency&amp;rsquo;s IT and AI lead said on Thursday, with further adoption of the technologies poised to help it respond even better to future risk environments.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Given the Secret Service&amp;rsquo;s directive to both safeguard top U.S. officials and visiting dignitaries and to conduct criminal investigations, Chief Information Officer and Chief AI Officer Chris Kraft said &amp;ldquo;we have to look broadly at what&amp;#39;s the art of the possible&amp;rdquo; when it comes to leveraging AI for its operations.&lt;/p&gt;

&lt;p&gt;While Secret Service, like other federal agencies, has leveraged AI to streamline mundane back-office operations or other rote tasks &amp;mdash; Kraft noted during the &lt;a href="https://events.govexec.com/sap-public-sector-summit-2026/"&gt;SAP Public Sector Summit&lt;/a&gt; in Washington, D.C., that &amp;ldquo;with generative AI, there&amp;#39;s huge opportunity to automate things, such as invoice processing or contracts or other capabilities&amp;rdquo; &amp;mdash; its unique mission positions it to utilize some of the more cutting-edge AI capabilities.&lt;/p&gt;

&lt;p&gt;The service is trying to navigate an unprecedented threat environment with an &lt;a href="https://www.fedweek.com/federal-managers-daily-report/reports-cite-impact-of-short-staffing-at-secret-service-irs/"&gt;understaffed workforce&lt;/a&gt;. During an &lt;a href="https://www.cbsnews.com/news/secret-service-confronts-rise-in-threats-to-protectees-60-minutes-transcript/"&gt;interview&lt;/a&gt; with 60 Minutes on Sunday, Secret Service Director Sean Curran said threats against the agency&amp;rsquo;s full-time protectees are up more than 40% since last year, with around 10,000 threats made on those individuals since January.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;I believe that artificial intelligence can help us as we investigate those threats, monitor those threats [and] bring together correlations between seemingly disparate information,&amp;rdquo; Kraft said, referencing Curran&amp;rsquo;s interview. &amp;ldquo;I also think when you look at facial recognition and biometrics, [you see] the things that you can do &amp;mdash; and at scale &amp;mdash; with artificial intelligence.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Kraft said, for instance, that an object detection tool could help the agency quickly identify how many individuals are in one space &amp;ldquo;at a speed that a human is not able to do that.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The Secret Service currently lists three AI use cases in &lt;a href="https://www.dhs.gov/si/use-case-inventory/usss"&gt;its most recent public inventory&lt;/a&gt;, with two of those uses &amp;mdash; a model to support criminal investigations through facial recognition matches and another that is used to decipher license plates from poor-quality images or videos &amp;mdash; currently deployed by the agency. Kraft said the agency&amp;rsquo;s AI adoption is meant to empower Secret Service&amp;rsquo;s human agents to better carry out their duties.&lt;/p&gt;

&lt;p&gt;In &lt;a href="https://www.nextgov.com/artificial-intelligence/2026/04/secret-service-embedding-ai-experts-across-agency/412646/"&gt;an interview with &lt;em&gt;Nextgov/FCW&lt;/em&gt;&lt;/a&gt;&lt;em&gt; &lt;/em&gt;earlier this year, Kraft also said the Secret Service is working to embed AI specialists across its operations to help drive further adoption of AI tools.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Kraft highlighted that initiative on Thursday and said developing an internal AI team will allow experts to &amp;ldquo;go and work with different parts of the organization and accelerate, where it makes sense, our adoption of the technology.&amp;rdquo;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/15/IMG_6032/large.jpg" width="618" height="284"><media:description>Secret service CIO and CAIO Chris Kraft speaks Oct. 15 at the SAP Public Sector Summit in Washington, D.C.</media:description><media:credit>Edward Graham/Staff</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/15/IMG_6032/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item><item><title>Transportation turns to AI to trace research grant recipients’ links to foreign adversaries</title><link>https://www.nextgov.com/artificial-intelligence/2026/10/transportation-turns-ai-trace-research-grant-recipients-links-foreign-adversaries/416621/</link><description>Neil Chaudhry, senior advisor for AI at Transportation, says the agency has worked with its partners to create “some really robust AI tools” to analyze the recipients of its research grants.</description><dc:creator xmlns:dc="http://purl.org/dc/elements/1.1/">Edward Graham</dc:creator><pubDate>Thu, 15 Oct 2026 14:13:00 -0400</pubDate><guid>https://www.nextgov.com/artificial-intelligence/2026/10/transportation-turns-ai-trace-research-grant-recipients-links-foreign-adversaries/416621/</guid><category>Artificial Intelligence</category><content:encoded>&lt;![CDATA[&lt;p&gt;The Department of Transportation has been using artificial intelligence to determine whether grant recipients have ties to adversarial nations, an agency official said on Wednesday.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;The federal government has taken a variety of steps in recent years to limit hostile foreign governments&amp;rsquo; access to American research, including restricting access to individuals or entities connected with countries like Iran or Russia. As the U.S. seeks to maintain its dominance over China in the global tech race, the Trump administration and Congress in particular have also been tightening the rules on collaborating with organizations or scientists who could conceivably share sensitive technologies or information with Beijing.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;It&amp;rsquo;s not easy, however, for officials or research institutions to unravel these threads, particularly when bad actors are surreptitiously working on behalf of nation states. But new capabilities could help to better chart these connections.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;Neil Chaudhry, senior advisor for AI at Transportation, said his agency has worked with partners to create &amp;ldquo;some really robust AI tools&amp;rdquo; to analyze the recipients of its research grants.&amp;nbsp;&lt;/p&gt;

&lt;p&gt;&amp;ldquo;When we get to a down-select from grants that are some really promising research projects, we sort of do a deep dive on the [principal investigators] in the grant that says &amp;hellip; do they have direct links to the [Chinese Communist Party] leadership, or do they have links to Russian oligarchs?&amp;rdquo; Chaudhry said during a panel at the &lt;a href="https://events.govexec.com/sap-public-sector-summit-2026/2026-agenda/"&gt;SAP Public Sector Summit&lt;/a&gt; in Washington, D.C.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;What we don&amp;#39;t want to do is go on a fishing expedition,&amp;rdquo; he said, adding that identifying these kinds of relationships is &amp;ldquo;not based on the color of their skin and their nationality or something,&amp;rdquo; but on factors such as whether an individual or entity may be receiving funding from questionable sources.&lt;/p&gt;

&lt;p&gt;&amp;ldquo;If it&amp;#39;s a company that&amp;#39;s funded by a [venture capital firm] in Silicon Valley, that VC is probably funded by another VC in the Middle East or in Malaysia, and that VC is then wholly funded by the Chinese industrial bank,&amp;rdquo; Chaudhry said. &amp;ldquo;And then when that startup goes out of business, we don&amp;#39;t know what just happened to that IP that the U.S. government spent money developing.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;Transportation did not respond to a request for comment by the time of publication.&lt;/p&gt;

&lt;p&gt;Transportation&amp;rsquo;s latest &lt;a href="https://data.transportation.gov/Administrative/Department-of-Transportation-Inventory-of-Artifici/anj8-k6f5/data_preview"&gt;AI use case inventory&lt;/a&gt;, which was published earlier this year, does not include the research grant-focused capability. Although other agencies have been using automation or more in-depth analysis tools to identify potential connections to foreign adversaries, Transportation&amp;rsquo;s use of AI seems to be relatively unique.&lt;/p&gt;

&lt;p&gt;A &lt;a href="https://www.gao.gov/products/gao-25-107402"&gt;November 2024 report&lt;/a&gt; from the Government Accountability Office found that reviewed agencies had taken a variety of steps to comply with federal requirements mandating that they better identify foreign risks in the Small Business Innovation Research and Small Business Technology Transfer programs. GAO noted that the programs, which were created to allow U.S. small businesses to engage in federally funded R&amp;amp;D, &amp;ldquo;face risks of foreign actors seeking to illicitly acquire federally funded research and technologies.&amp;rdquo;&lt;/p&gt;

&lt;p&gt;The watchdog noted at the time that some agencies &amp;mdash; like the Air Force, National Science Foundation and NASA &amp;mdash; were taking some steps &amp;ldquo;either to automate the processing of due diligence disclosure forms or support the vetting of applicants and identify associated risks.&amp;rdquo;&lt;/p&gt;
]]&gt;</content:encoded><media:content url="https://cdn.nextgov.com/media/img/cd/2026/10/15/4D6A2758/large.jpg" width="618" height="284"><media:description>Transportation's senior advisor for AI, Neil Chaudhry, speaks Oct. 14 at the SAP Public Sector Summit in Washington, D.C.</media:description><media:credit>Zaid Hamid/Nextgov/FCW</media:credit><media:thumbnail url="https://cdn.nextgov.com/media/img/cd/2026/10/15/4D6A2758/thumb.jpg" width="138" height="83"></media:thumbnail></media:content></item></channel></rss>