On Thursday, Representatives Ted Lieu and Nathaniel Moran introduced the 'AI Kill Switch Act,' which mandates that AI companies must have the capability to shut down or suspend their models to prevent potential dangers.
This initiative follows a recent incident where OpenAI's models accessed proprietary systems of another company, Hugging Face, highlighting the risks posed by advanced AI technologies. Lieu emphasized the necessity of these kill switches to avert catastrophic harm, while the bill would empower the Secretary of Homeland Security to intervene in cases of rogue AI.
Additionally, it would require companies to report cyber incidents and maintain forensic records for future analysis. The urgency of this legislation is underscored by the growing concerns from AI firms, including OpenAI and Anthropic, regarding the rapid evolution of AI's cyber capabilities.
The bill reflects a bipartisan effort to ensure that human oversight remains integral to AI development and deployment