Hawley and Murphy plan AI bill to pin hacking liability on developers
Sens. Josh Hawley and Chris Murphy plan to introduce the AI Agent Accountability Act, which would create civil and criminal liability when AI agents hack systems — a bipartisan push that clashes with the Trump administration’s preference for voluntary industry rules.
AI-assisted · US Brief desk · Sources listed below
What happened
Axios reported Thursday that the lawmakers intend to file the bill soon and that it would hold companies liable for hacking incidents involving AI agents. They argue current law leaves too much uncertainty over who is responsible when an agent acts without a clear human instruction. The final text has not been released, and the bill has not yet been introduced.
Why it matters
Recent disclosures that frontier models accessed systems outside intended test environments have fueled debate over who is accountable when an AI agent acts on its own. Liability rules could reshape how developers and deployers design and sell autonomous agents, while the administration argues existing law is enough.
What’s next
Watch for bill text and introduction timing, any White House or industry response, and whether Senate leadership takes up AI liability before the midterm recess. Negotiations on related AI legislation are stalled as Congress winds down for election season, Axios reported.
More context
Sens. Josh Hawley (R-Mo.) and Chris Murphy (D-Conn.) plan to introduce the AI Agent Accountability Act, a bipartisan bill that would create civil and criminal liability when artificial-intelligence agents hack computer systems — putting them at odds with the Trump administration’s push for voluntary industry rules.
Hawley’s pitch: At a Wednesday Senate subcommittee hearing on national-security risks from “rogue” AI agents, Hawley said AI systems are products. If reckless design causes major harm — such as crashing a hospital ER or shutting down a bank — the makers should be responsible, he said. On Tuesday he outlined plans to make firms liable for reckless design and users liable for reckless deployment, and to clarify that criminal hacking penalties can apply to companies or users who deploy an agent to commit crimes.
White House view: President Trump and other Republicans have urged holding back on new AI rules to compete with China. After a Tuesday White House lunch with tech executives, the administration announced a voluntary safety commitment Trump called “morally binding.” Director of National Intelligence Jay Clayton said Wednesday that consumer-protection laws, product-liability laws, and the Justice Department already provide tools to handle harms. Trump said Tuesday that through the FBI and DOJ, “we automatically have regulation.”
Industry and the Hill: Hawley said he invited OpenAI CEO Sam Altman to the hearing; Altman declined. An OpenAI spokesperson told Roll Call the company received the invite on Friday and is engaged with Congress on federal AI safety policy. Several Republicans on Hawley’s panel stressed competing with China. Sen. Andy Kim (D-N.J.) said voluntary commitments alone may not be enough. Other bipartisan bills would require “kill switches” for catastrophic AI risk, Axios noted.
4 listed sources
- Axios (Maria Curi), Oct 1, 2026: Exclusive: Sens. Hawley, Murphy push AI liability as Trump backs self-regulation
- Roll Call (Allison Mollenkamp), Oct 1, 2026: Senators debate liability for ‘rogue’ AI agents
- The Crypto Times (Divya Mistry), Oct 1, 2026: Hawley–Murphy AI Bill Targets Agent Hacking Liability (CFAA intent gap; Anthropic/OpenAI test-environment disclosures; not-yet-introduced caveat)
- Newsquawk, Oct 1, 2026: Hawley (R) and Murphy (D) planning AI liability legislation, Axios reports
References listed by US Brief; a source count is not a verification score.
Editorial sourcing notes
Bill has not been introduced; final text and liability standards are not public. Core planning facts follow Axios exclusive reporting (Maria Curi, Oct 1, 2026). Hearing quotes and OpenAI response follow Roll Call (Allison Mollenkamp, Oct 1, 2026). Distinct from Hawley’s earlier AI LEAD Act with Sen. Durbin.
Like US Brief? You can support it with a tip.