Key facts
- Anthropic proposed courts should have power to block AI programs posing serious risks.
- AI companies should develop kill switches and embed outside safety auditors, Anthropic said.
- The company suggested independent evaluators report risks to the state Attorney General.
- The Attorney General could seek a court order to suspend a model's deployment if it poses a near-term risk of catastrophic harm.
Anthropic has outlined its vision for AI regulation, urging that courts should possess the authority to block AI programs deemed to pose significant risks. The artificial intelligence company also advocates for AI firms to implement 'kill switches' for their technologies and to integrate independent safety auditors within their operations. These policy recommendations were detailed in a letter sent to California Governor Gavin Newsom earlier this month, amidst growing concerns over advanced AI capabilities and potential misuse.