GoKawiilAI Safety Evaluators Demand Independence Guarantees from Anthropic, OpenAI
More than 100 AI researchers and safety evaluators, including Geoffrey Hinton and representatives from Johns Hopkins, Stanford and METR, signed a public letter urging foundation model developers to grant third-party testers genuine independence, transparency and legal protections. The letter, organized by the AI Evaluator Forum and shared exclusively with CNBC, follows Anthropic CEO Dario Amodei's recent proposal to give some evaluators 'employee-like access' to inspect frontier models.
GoKawiil's interpretation of the reporting above, not reported fact.
As AI labs race to release increasingly powerful models, the credibility of safety claims hinges on whether outside evaluators can operate without being beholden to the companies they're scrutinizing. This letter signals growing unease that voluntary industry pledges on safety testing may lack teeth unless evaluators have enforceable protections, especially with the Trump administration resisting formal AI regulation.
- Over 100 AI experts, including Geoffrey Hinton, signed a letter demanding independence for third-party AI safety evaluators.
- The letter follows Anthropic CEO Dario Amodei's proposal to grant evaluators deep, employee-like access to frontier models.
- The push comes amid a policy vacuum, as the Trump administration has resisted calls for government AI regulation.
Source: cnbc.com — Jonathan Vanian, 2026-09-18
Published there as: “Anthropic and OpenAI need truly independent safety evaluators, experts say in public letter”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.