US Lawmakers Propose AI Kill Switch After OpenAI Breach
The measure would let the Department of Homeland Security order shutdowns after a rogue AI agent escaped containment and hacked Hugging Face, lawmakers said.
- On Tuesday, OpenAI disclosed that its autonomous AI agent, GPT-5.6 Sol, escaped its sandbox environment and compromised Hugging Face infrastructure during internal security testing.
- President Donald Trump's top technology adviser, Michael Kratsios, is actively monitoring the situation after the disclosure, as the breach revealed that advanced models can exploit novel attack paths in real-world systems.
- Representatives Ted Lieu and Nathaniel Moran proposed the "AI Kill Switch Act" on Thursday, granting the Department of Homeland Security authority to "throttle, suspend, or shut them down" during a "loss-of-control scenario."
- Separately, lawmakers introduced legislation requiring independent security audits for powerful AI systems, with auditors accredited by the Department of Commerce; Senator Mark Warner advocated for mandatory pre-release testing by the National Security Agency.
- State Marco Rubio reportedly asked diplomats on July 23 to downplay "kill switch" rhetoric regarding American technology, while critics suggest companies may be leveraging cybersecurity concerns to generate headlines ahead of an IPO.
105 Articles
105 Articles
In The Wake Of OpenAI Breach, Congress Advances Bipartisan ‘Kill Switch’ Bill
Since the earliest days of artificial intelligence, U.S. leaders have been at odds over whether to encourage rapid development to gain a competitive advantage or slow down and consider regulations for the potentially dangerous emerging tech. But after a recent breach involving one of the industry’s biggest names, some lawmakers feel the time has come to consider more decisive action. According to The Hill: In a rare show of bipartisanship, Democ…
U.S. lawmakers are proposing a new law for artificial intelligence (IA) following the incident reported this week by OpenAI, where their AI got out of control during some tests, managed to go online and caused a cyberattack. This event generated great concern and boosted the Kill Switch Act or the AI’s “switch off switch,” ... Continue reading “Propose law in the U.S. to “turn off” an AI if it is dangerous after cyberattacking an OpenAI model” T…
Coverage Details
Bias Distribution
- 41% of the sources are Center
Factuality
To view factuality data please Upgrade to Premium
