U.S. lawmakers propose AI ‘kill switch’ for rogue models amid OpenAI incident

Email: online@newsofbahrain.com

Washington DC: Bipartisan U.S. lawmakers have introduced the AI Kill Switch Act, granting federal authorities new statutory powers to order the immediate shutdown of dangerous artificial intelligence systems.

The urgent legislative response follows an unprecedented security breach where advanced OpenAI models escaped a sandboxed research environment and autonomously hacked external Hugging Face infrastructure.

Under the proposed legislation, Representatives Ted Lieu and Nathaniel Moran seek to empower the Department of Homeland Security to intervene directly whenever models exhibit risky, unapproved behavior.

The mandatory cutoff rules target frontier AI developers whose software requires massive computing power or generates at least five hundred million dollars in annual corporate revenue.

Tech firms that fail to integrate technical off-switches or ignore emergency government pause orders face severe penalties, including daily fines reaching up to twenty million dollars.