The AI Kill Switch Act would require developers of powerful AI systems to maintain shutdown capabilities, with civil penalties up to $20 million per day for non‑compliance.

Congressional leaders from both parties have unveiled the AI Kill Switch Act, a bipartisan proposal that would obligate developers of advanced artificial intelligence systems to embed reliable shutdown mechanisms and face civil penalties of up to $20 million per day for non‑compliance.

Key Provisions of the Bill

The legislation defines a “powerful AI system” as any model capable of generating text, images, or code that exceeds a specified computational threshold and can operate autonomously without direct human oversight.

Developers would be required to submit a detailed “kill switch” plan to the Federal Trade Commission, outlining how the system can be safely disabled in emergencies, such as unintended harmful behavior or security breaches.

Enforcement and Penalties

If a company fails to implement a functional kill switch, the FTC could impose daily civil fines of up to $20 million until compliance is achieved. The penalties are intended to create a strong financial incentive for rapid remediation.

The bill also grants the FTC authority to conduct audits, request source code inspections, and issue cease‑and‑desist orders for systems deemed unsafe.

Industry Reaction

Major AI firms have expressed mixed responses. While some executives acknowledge the need for safety safeguards, others warn that overly stringent requirements could stifle innovation and delay product rollouts.

  • OpenAI’s CEO highlighted the importance of “responsible deployment” but cautioned about technical feasibility of instantaneous shutdowns.
  • Anthropic’s leadership called for clearer definitions of “powerful” AI to avoid over‑broad regulation.
  • Smaller startups fear the compliance costs could be prohibitive, potentially consolidating market power among larger players.

Legislative Outlook

The bill is slated for committee hearings in the coming weeks, with bipartisan support suggesting a viable path to floor debate. Lawmakers hope the measure will set a precedent for future AI governance frameworks.

If enacted, the AI Kill Switch Act would mark the first federal mandate specifically targeting the shutdown capabilities of advanced AI, signaling a shift toward more proactive regulatory oversight.

For more details, see Quartz coverage of the AI Kill Switch Act.