Anthropic pledges permanent system access for outside AI evaluators

Dario Amodei is calling for the AI industry to slow down, describing a three-part plan in his essay “We Must Pace the Frontier.” He says Anthropic is committing to the first step: permanent, employee-level access to its systems for third-party evaluators to verify adherence to safety measures, report incidents and assess models’ alignment during training.

AVERI supports embedding independent experts within frontier AI companies and wants to move beyond voluntary auditing toward binding requirements. It says that requires standards for auditor qualifications and independence, audit methods and reporting, and the safety and security standards companies are audited against. AVERI says it is conducting pilot projects to help inform audit standards.