Anthropic co-founders urge mandatory safety oversight as AI races ahead

Dario Amodei and Jack Clark have called for slowing AI development and imposing mandatory, verifiable safety controls, as policymakers weigh regulation.

• Amodei warned that the pace of AI development may outstrip the ability of companies and governments to install adequate safety measures.
• Amodei proposed independent safety monitors, common standards, capability checkpoints and international coordination, calling a full halt to AI development unrealistic.
• Clark said a third-party-verifiable "kill switch" may need to be mandatory for AI companies under future rules.
• US lawmakers introduced the Kill Switch Act, requiring companies to build shutdown mechanisms into AI tools.
• UK First Secretary Louise Haigh said the government must heed warnings from AI industry leaders and may need to update its regulatory framework.
• OpenAI's Tom Duff Gordon endorsed stronger UK AI rules but urged regulators to focus on the most powerful developers rather than startups.