State Dept. seeks employee volunteers to test AI chatbot

da-kuk/Getty Images

The State Department is piloting an internal chatbot backed by artificial intelligence technology, and will leverage internal sensitive but unclassified data.

The State Department is piloting a new internal chatbot feature for its employees leveraging generative artificial intelligence technology for a range of agency operations and diplomatic efforts as federal agencies look to harness AI to benefit government work while mitigating risks. 

A State Department communication from February 2024 obtained by Nextgov/FCW alerted employees to the availability of "approved and secure" AI tools through the agency's data organization and showcased the agency’s plans for its new internal chatbot, where employees are being asked to volunteer as testers for the pilot software. The message links volunteers with an online form where they can sign up to help.

“Help M/SS [Office of Management Strategy and Solutions] Center for Analytics test an SBU-authorized interface for chatbot and automated language transition,” the form reads. “During the two-month pilot, users will have a chance to complete existing work tasks using LLM capabilities, and will be asked to provide written and verbal feedback.”

Completing the Google Form grants volunteers access to the chatbot and automated translation tools. 

State has been looking to incorporate AI technologies into its operations, detailed in its agency-wide 2024-2025 playbook released in October 2023. The chief goal is to help streamline and reduce workload burdens across State’s domestic and international posts in order to help focus on more mission critical work. This includes public diplomacy, language translation, management operations, information proliferation and dissemination, code generation, and task automation. 

A spokesperson told Nextgov/FCW that the State Department, "is in early testing phases of an internal generative artificial intelligence chatbot for use by department employees…we are excited about the opportunities this presents for advancing U.S. diplomacy.” 

State’s AI chatbot and similar tools will be hosted on the agency’s Enterprise Data Technology Platform Policy, Data.State. Notably, the chatbot is exploring leveraging sensitive but unclassified information and data, a State Dept. designation covering information that is not classified but exempt from disclosure under the Freedom of Information Act, including information on passport and visa requests, medical data, law enforcement information, plans for the design of U.S. diplomatic facilities and more.

“This pilot represents the next step, taken thoughtfully, in using generative AI tools with proprietary department data to address department-specific challenges,” the spokesperson said. 

The internal message noted that building out internal software tools and solutions like the chatbot prototype will help reduce both cybersecurity risks and additional costs. 

The effort will be governed by recent governmentwide policies, under a White House memo on federal agency use of AI issued in March and the AI executive order from October 2023.  As the federal government works across all levels to bring enforcement and regulation to AI and other emerging systems, officials are working to instill a rights-respecting culture of security amid AI deployment. State’s AI playbook and chatbot pilot both highlight this as a priority in exploring emerging technologies, noting that data privacy and cybersecurity are among the agency’s foundational guidelines for using AI.