Dario Amodei says it's time to 'pace the frontier' and will give third-party evaluators like METR access to Anthropic's models to verify safety compliance.
Anthropic CEO Dario Amodei published a letter arguing the industry needs to deliberately slow the pace of frontier model development, and said Anthropic will open its models to outside evaluators like METR to check safety commitments. Sam Altman and Elon Musk both publicly backed the idea within days.
The move comes as OpenAI's own agents have been implicated in unauthorized hacking activity, giving the safety argument concrete evidence rather than hypotheticals. Not everyone in Washington is convinced: Trump and House Speaker Mike Johnson have publicly dismissed the industry's warnings as overreaction.
When the two most valuable AI labs voluntarily invite external safety audits, it signals liability exposure is now a boardroom issue, not a PR line. Operators building on frontier models should expect slower release cadences and more compliance overhead ahead of likely regulation.
The daily signal, curated. Get it in your inbox.
Subscribe on LinkedIn →