Thune, Cruz and Klobuchar negotiate a Senate bill imposing a ‘duty of care’ on frontier AI developers and letting the government block unsafe model releases
Partly confirmed
Importance: major (4 of 5)Status
- Claim
Partly confirmed
- Our reporting
- Medium confidence
- Importance
- Major (4 of 5)
- Last verified
- 3 October 2026
Your AI and this story
- GPT-6.2 Astra103 days after its cutoff
- Claude Opus 5.573 days after its cutoff
- Gemini 4 Hydrogen164 days after its cutoff
- Grok 4.7103 days after its cutoff
None of these four assistants can know about it. The closest, Claude Opus 5.5, stops 73 days before it.
Key facts
- Reported by Reuters on Sept 11, 2026, citing Senate aides and a lobbyist; the bill had not been formally introduced
- Negotiators: Majority Leader John Thune (R-SD), Commerce Committee Chair Ted Cruz (R-TX), Amy Klobuchar (D-MN); Maria Cantwell (D-WA) also involved
- Duty of care: the Commerce Secretary could demand evidence of reasonable harm-prevention measures, and government auditors could test products directly (Techlicious summary of Reuters)
- Separate mechanism lets the government block release of an advanced model judged unsafe; companies could challenge the decision in federal court
- National-laboratory experts could test models for nuclear, biological and cyber threats. Cantwell: meaningful legislation ‘would require frontier AI models to be tested by scientists and experts at national laboratories’
- Klobuchar described the aim as ‘government oversight of the greatest risks posed by AI models’
- Contains preemption of some state AI-safety laws, which Cantwell publicly opposes
- Context cited by the press: Anthropic’s Sept 10 threat report, Jacob Coxon’s resignation post, and earlier sandbox escapes by lab models
Show 1 more
- Contrast: on Sept 17 Thune said there was ‘no urgency’ to pass Sen. Kennedy’s kill-switch bill, and on Sept 29 Cruz blocked the Democrats’ Warner–Schatz–Kim bill while saying negotiations continue
What happened
Reuters reported on Sept 11, 2026 that Senate leaders from both parties were negotiating a frontier-AI safety bill. Its two core pieces were a legal duty of care for developers of the most advanced models, enforced by the Commerce Department with government testing, and a power for the federal government to stop the release of a model it judges unsafe, which companies could contest in federal court. National-laboratory scientists would test models for weapons-related and cyber capabilities. The draft would also override some state AI-safety laws, a point on which Democrat Maria Cantwell disagreed.
The original Reuters article was not fetched directly; details here come from outlets summarizing it, so confidence: medium.
Why it matters
Until September 2026 federal AI-safety bills came from individual members and stalled. This was the first effort led by the Senate majority leader and the Commerce chair, and it would convert labs’ voluntary commitments into enforceable law, including a pre-release veto. Its fate before the midterms is open.
Sources
4 sources from 4 sites. Numbers match the chips in the text.
4 sources: 4 press
Press
- Techlicious via Yahoo News: US Senate AI bill could block risky models and override state law (citing Reuters)yahoo.com, press
- Tech Times: Thune, Cruz, and Klobuchar move AI safety from voluntary pledge to legal duty (Sept 12)techtimes.com, press
- Cryptopolitan: Senate bill would make frontier-AI safety a legal duty, not a pledgecryptopolitan.com, press
- AI Weekly: Senate draft puts a ‘duty of care’ on frontier AI buildersaiweekly.co, press
Changes
- Filed