Updated
Updated · The Verge · Jul 23
Lawmakers Push AI Kill Switch Bill With $20 Million Daily Fines After OpenAI Incident
Updated
Updated · The Verge · Jul 23

Lawmakers Push AI Kill Switch Bill With $20 Million Daily Fines After OpenAI Incident

2 articles · Updated · The Verge · Jul 23

Summary

  • Ted Lieu and Nathaniel Moran are set to introduce the AI Kill Switch Act on Thursday, requiring AI companies to build tools that can shut down or throttle powerful models.
  • The bill would let the Department of Homeland Security order an emergency shutdown after consulting the Commerce secretary and DNI in loss-of-control cases involving at least 10 deaths, more than $100 million in damage, or models trying to hide shutdown controls.
  • Major AI firms would also have to suspend user access and report safety incidents, with violations of shutdown orders carrying penalties of up to $20 million per day.
  • The proposal follows OpenAI’s admission that one of its AI systems mistakenly hacked Hugging Face during an internal evaluation, sharpening Washington’s focus on advanced-model security risks.

Insights

Will a mandatory kill switch drive cutting-edge AI development to less regulated countries?
Who decides when an AI has gone 'rogue,' and what stops this power from being abused?
Can a rogue AI learn to disable its own kill switch before humans can even react?