Members of Congress put forward a cluster of artificial intelligence proposals this week that would make AI developers legally responsible for harm their systems cause, require federal agencies to tell people when AI shapes decisions about them and give the Cybersecurity and Infrastructure Security Agency (CISA) a dedicated AI security unit, Nextgov/FCW reported Friday. All are at an early stage, but together they sketch where accountability rules could land for companies building or deploying AI.

What Happened

Developer liability. Rep. Lori Trahan, D-Mass., released a discussion draft of the Clear Liability for Artificial Intelligence Misconduct (CLAIM) Act on Oct. 7. Developers would be liable when an AI system injures a third party through conduct that would be negligence, an intentional tort or a crime if a person did it, unless the user or a company that modified the system intended the conduct or was negligent. Courts would presume the system had the state of mind a person in the same position would have had, so developers could not argue that AI cannot form intent. Injured people could sue under a new federal cause of action; the draft sets a floor that leaves state claims and state AI laws in place. Trahan's office is taking comments.

Sens. Josh Hawley, R-Mo., and Chris Murphy, D-Conn., have separately proposed holding developers and operators of AI agents, software that carries out tasks on its own, criminally and civilly liable for hacking incidents, Nextgov/FCW reported.

RelatedWhite House AI Task Force Plans Breach Warnings From Labs at Its First Meeting

Agency disclosure. Sens. Gary Peters, D-Mich., and James Lankford, R-Okla., introduced a bipartisan bill requiring agencies to notify people when they interact with AI or automated systems, and when an agency makes a critical decision about them using an "AI-augmented decision process," Federal News Network reported. The Office of Management and Budget would issue guidance, and people who receive an adverse decision would get a redress process with human review.

A CISA task force. Reps. Troy Carter, D-La., and Bennie Thompson, D-Miss., the top Democrat on the House Homeland Security Committee, reintroduced the CISA Securing AI Task Force Act on Oct. 8. Within a year of enactment, CISA's director would set up a task force to coordinate the agency's AI work, advise stakeholders on cyber risks from deploying AI and identify workforce gaps, briefing the House and Senate homeland security committees twice a year. Similar legislation was filed in May 2024, Nextgov/FCW noted.

A White House AI office. Rep. Tom Barrett of Michigan introduced the Artificial Intelligence Governance and Emerging Threats Act on Oct. 9. It would create an Office of Artificial Intelligence and Emerging Technology Threats in the Executive Office of the President, modeled on the Office of the Director of National Intelligence, led by a Senate-confirmed, Cabinet-level official with five confirmed deputies who would testify regularly to Congress.

Sen. Maria Cantwell, D-Wash., the top Democrat on the Senate Commerce Committee, released a framework under which the National Institute of Standards and Technology (NIST) would set standards for frontier models, the most capable AI systems, and covered models could not be released until an independent audit confirmed compliance, Federal News Network reported.

Why It Matters

For AI developers, the Trahan draft and the Hawley-Murphy proposal point toward the same exposure: harm done by an AI system, including an autonomous agent, treated as the developer's legal responsibility. Barring the argument that AI lacks intent removes a likely defense, and because the draft would not override state law, companies could face federal and state claims at once. Cantwell's audit model would add a compliance gate before launch. Vendors supporting federal decisions would feel the Peters-Lankford bill, which requires disclosure and human review.

The CISA and White House office bills address who in government owns AI security. "This bill creates a dedicated CISA task force so the agency can better understand the risks that come with AI," Thompson said in a statement. Federal News Network noted that neither CISA nor NIST is listed as a member of the administration's new "Super Intelligence" task force.

Legislative chamber with rows of red seats seen from above
Illustrative photo: Unsplash. Not a photograph of the reported event.

What to Watch

  • The CLAIM draft. Whether Trahan formally introduces it.
  • Committee action. Whether any bill gets a hearing or markup in the weeks before the midterm elections.
  • Cantwell's framework. Whether it becomes bill text.
  • Agency roles. Whether CISA and NIST gain a formal place in the administration's AI efforts.