Congress wants to set the stage for AI regs

A new bill would establish a federal advisory committee to examine the impact of AI.

Shutterstock image (by razum): fusion of human and artificial intelligence.

A bipartisan group in Congress wants to build a resource to guide policymakers as they confront the mammoth challenge of regulating artificial intelligence.

The FUTURE of AI Act would establish a federal advisory committee to examine a wide range of questions surrounding how technologies like automation, machine learning and other forms of AI would impact society. The legislation would also empower the committee with making recommendations to the secretary regarding implementation and regulation in the public and private sector.

The bill was introduced in the Senate by Sens. Maria Cantwell (D-Wash.), Ed Markey (D-Mass.), and Todd Young (R-Ind.). A companion bill in the House is sponsored by Reps. John Delaney (D-M.D.) and Pete Olsen (R-Texas.).

Delaney, co-chair of the House AI caucus along with Olsen, framed the committee as a body that would map out the multitude of ways that AI technologies are expected to affect society and help to guide both public and private sector implementation.

"AI is going to reshape our economy the way the steam engine, the transistor or the personal computer did and as a former entrepreneur, I believe the impact will be positive overall," said Delaney. However, he added, "big disruptions also create new policy needs and we should start working now so that AI is harnessed in a way that society benefits, that businesses benefit and workers benefit."

The advisory committee would be composed of 19 voting members pulled from industry, academia, labor and civil rights groups. Representatives from federal agencies like the National Institute for Standards and Technology, the National Science Foundation, the Department of Justice and the Federal Trade Commission will also have seats as non-voting members.

Members would be charged with looking into the pros and cons around how AI technologies will impact the workforce, international cooperation and competitiveness, open data sharing, and legal rights, among other issues. Additionally, it would investigate how to promote public and private sector investment; how human bias can be transmitted to algorithms to affect the outcome of software-based decision-making; and how best to work through ethics and standards for both industry and government.

Libertarian think tank R-Street was among the organizations that consulted with members of Congress while crafting the bill. Technology policy associate Caleb Watney told FCW his organization advised the group to stay away from establishing a commission with standalone rulemaking authority, largely because the technologies are so nascent that any attempt to institute hard and fast regulations at this stage would likely backfire.

"We have so little knowledge about the way [AI] might develop, the timelines, the types of technologies it might lead to, that the likelihood that…any specific regulatory action we take today having its intended effect is very, very small," said Watney.

That's in part because Watney believes there are still many federal agencies and members of Congress who do not fully understand the different capabilities and applications of AI. For example, he said government should be more tightly regulated in how it purchases and uses artificial intelligence than the private sector, largely because the state has domain over the criminal justice system and other public goods that require a higher level of transparency and trust from the public. An AI algorithm that has bias coded in can have dramatically starker consequences when it is tied to a system that decides whether to audit your taxes or deny you bail.

Watney also worries that moving too quickly might result in what he calls "the global arbitrage" effect, where a ban on certain AI applications in the U.S. or European Union winds up pushing development to other world powers looking for an edge in the AI arms race. China, for example, has set a goal to become the world’s leading incubator for artificial intelligence by 2025.