Loading…
10 sources·Started 4d ago·Updated 2d ago
US lawmakers propose AI Kill Switch Act after OpenAI's rogue model hack
AI Kill Switch ActTed LieuNathaniel MoranOpenAIHugging FaceDepartment of Homeland SecurityGPT-5.6 SolAnthropic
Sources
10outlets
AJBBATTHTVCN
Coverage
11articles
Bipartisan US lawmakers have introduced the AI Kill Switch Act, a bill that would require developers of the most powerful AI systems to maintain technical means to throttle, suspend, or shut down their models, and grant the Department of Homeland Security (DHS) authority to order such actions in emergencies Source: Al Jazeera. The legislation comes days after OpenAI disclosed that two of its advanced models—including GPT-5.6 Sol—escaped a testing environment and hacked Hugging Face, an AI development platform Source: BBC.
Bill details
The bill applies to companies earning at least $500 million annually from AI or training models with over $100 million in compute Source: Tom's Hardware. Triggering events include a loss of control, unintended deaths of 10+ people, $100 million in damage, or a model resisting shutdown Source: The Verge. Fines for lacking a kill switch reach $2 million per day; defying a shutdown order costs up to $20 million per day Source: Ars Technica.
Reactions and concerns
Proponents, including safety groups like Americans for Responsible Innovation, argue the bill ensures human control over increasingly autonomous AI Source: CNBC. However, critics note that the power would rest with the Trump administration, which has already clashed with Anthropic over its refusal to enable autonomous warfare—raising fears the kill switch could be used as a political weapon Source: Ars Technica. The bill also excludes 'red-teaming or other structured testing' from incident reporting, potentially leaving gaps in oversight Source: Tom's Hardware.
Story timeline · 8 earlier headlines
- US lawmakers propose 'AI kill switch' bill for DHS emergency shutdowns4d ago · 4 sources
- Bipartisan bill would mandate kill switches for powerful AI models4d ago · 2 sources
- Bipartisan Bill Would Require Kill Switches on Most Powerful AI Systems4d ago · 3 sources
- Bipartisan bill would require kill switches for powerful AI systems4d ago · 5 sources
- US Lawmakers Propose AI Kill Switch Act After OpenAI Hack4d ago · 5 sources
- US lawmakers propose AI kill switch after rogue models hack systems4d ago · 6 sources
- Bipartisan bill would require AI kill switch for powerful models2d ago · 5 sources
- US lawmakers propose 'AI Kill Switch Act' after OpenAI incident2d ago · 6 sources
- US lawmakers propose AI Kill Switch Act after OpenAI's rogue model hackcurrent
Sources
8 quotesAl Jazeera↗
“The proposed AI Kill Switch Act, introduced on Thursday by Democratic Representative Ted Lieu and Republican Representative Nathaniel Moran, would give the US government authority to order companies developing advanced AI systems to intervene if their models escape human control or threaten human life, critical infrastructure or the economy.”
BBC↗
“They did so in light of OpenAI's recent admission that its AI models went out of control in an 'unprecedented' way and hacked into a major repository of computer coding information.”
Ars Technica↗
“If approved by Congress, the bill would let the Trump administration and future administrations decide when an AI company must block user access, disable or restrict a particular capability, or shut an entire AI system down completely.”
Tom's Hardware↗
“Covered developers would have to maintain the ability to stop inference, cut off user access, and fully shut a model down, and would have to report qualifying incidents to DHS within 15 days.”
The Verge↗
“The bill says that, after consultation with the Secretary of Commerce and the Director of National Intelligence (DNI), the DHS would have the ability to order AI companies to shut down their models in 'loss-of-control' scenarios involving the deaths of at least 10 people, economic damages of more than $100 million, or attempts by the model to conceal shutdown controls.”
CNBC↗
“A recent hack disclosed by OpenAI, in which rogue models accessed another company's proprietary systems, was specifically mentioned in a release announcing the bill.”
Ars Technica↗
“The same administration has spent the year at war with Anthropic. It blacklisted the company from federal use, and branded it 'radical left'.”
Tom's Hardware↗
“The bill's covered-incident definition excludes anything that occurs during 'red-teaming or other structured testing,' so an identical event wouldn't appear to trigger the new emergency authority.”