H.R. 1694: AI Accountability Act

H.R. 1694, the AI Accountability Act, is not a regulation. It does not fine companies, ban anything, or set compliance rules. It is a fact-finding assignment: Congress is telling the executive branch to study how AI should be checked before it writes any binding rules.

The job goes to the National Telecommunications and Information Administration, the Commerce Department's telecom agency. The agency has to study "accountability measures" for AI and report to Congress within 18 months. The bill defines an accountability measure as an audit, an assessment, or a certification meant to give assurance that a system is trustworthy. That definition steers the conversation toward concrete oversight tools rather than vague promises from the companies building the technology.

The study is broad. It has to look at how those measures are being built into communications networks, including telecom systems and social media platforms, plus electromagnetic spectrum sharing. It also has to weigh whether AI accountability can help close the digital divide, expand digital access in the U.S., and cut risks like cybersecurity threats. And it has to untangle the language itself: how the word "trustworthy" gets used in AI, and how it relates to terms like "responsible" and "human-centric." The bill's drafters clearly think the AI debate is full of loose words and want sharper ones before tougher rules get written.

A second piece of the bill is about transparency. It requires public meetings with industry, academics, and consumers to gather input on what information people should be able to see about the AI systems they interact with, are affected by, or study, and how that information should be shared. That feedback feeds a second report, also due within 18 months, sent to the House Energy and Commerce Committee and the Senate Commerce Committee. Boiled down, Congress is asking two questions before it moves further: how should AI be checked, and what should the public be allowed to know about it?