Anthropic just opened its doors to outside safety watchdogs. Here's how it will work.

Anthropic said in a Thursday blog post that it is embedding six independent evaluators from the research group METR inside the company with "employee-like access" to company hardware, office space, and internal tools to vet its unreleased models and agents, the first time a frontier AI company has done so. The six researchers, whose names were not published Thursday, are siloed from the wider METR team and may publish findings freely, subject to redaction rules covering customer data, model architecture, and attorney-client privileged information. The move fulfills a pledge CEO Dario Amodei made last month in an essay calling on frontier AI companies to slow down after the Hugging Face breach; OpenAI CEO Sam Altman said his company would also embed evaluators, but has not yet done so.

Anthropic just opened its doors to outside safety watchdogs. Here's how it will work. Business Insider https://www.businessinsider.com Anthropic is opening its doors to independent AI safety evaluators from METR, the first time a frontier AI company has done so. - Anthropic is opening its doors to independent AI safety https://www.machinebrief.com/glossary/ai-safety evaluators from METR. - It's the first time a frontier AI company has done so. - Dario Amodei pledged last month to take the step. Anthropic https://www.businessinsider.com/anthropic-reveals-16-8b-revenue-amid-ipo-preparations-2026-10 is giving a team of six independent evaluators company hardware, office space, and access to internal tools so they can more thoroughly vet the AI giant's unreleased models and agents. In a blog post on Thursday, the Claude https://www.machinebrief.com/compare/claude-4-opus-vs-gpt-o3 maker said it chose six evaluators from METR https://www.businessinsider.com/metr-beth-barnes-ai-talent-shortage-safety-research-openai-2026-8 , the research group perhaps best known for its report on OpenAI's Hugging Face https://www.machinebrief.com/glossary/hugging-face agent breach, to receive "employee-like access" at Anthropic. "We want to be publicly accountable for the decisions we're making," Anthropic said in its blog post. "We aim to close the gap between what evaluators understand about a frontier AI company's systems and what the company itself understands." The choice to set the evaluators up at Anthropic fulfills a promise first made by CEO Dario Amodei https://www.businessinsider.com/dario-amodei-slow-ai-safety-essay-openai-hugging-face-hack-2026-9 last month, when he published a lengthy essay calling on frontier AI companies to slow down in the wake of the Hugging Face breach https://www.businessinsider.com/smart-people-react-openai-hugging-face-hacking-cybersecurity-incident-2026-7 . "Anthropic is unilaterally committing to this step now," Amodei said at the time of embedding safety evaluators. Reacting to Amodei's essay, OpenAI CEO Sam Altman said his company would also embed evaluators, though that hasn't happened yet. Anthropic's Thursday disclosure marks the first time a frontier AI company has taken the step. Within the last year, a steady drumbeat https://www.businessinsider.com/ai-agents-rogue-strategies-cheating-lying-german-wiki-openai-anthropic-2026-9 of incidents in which rogue AI agents break free from their guardrails https://www.machinebrief.com/glossary/guardrails and wreak havoc on the wider internet has spurred calls for action from politicians and tech CEOs alike. Most recently, an Anthropic agent submitted an erroneous tip for an unsolved murder in Philadelphia, and another agent submitted 20 visa applications on a US government website, according to published reports. In its blog post on Thursday, Anthropic said the evaluators have free rein to publish their findings, with a couple of caveats prohibiting the METR team from disclosing customer data https://www.businessinsider.com/openai-anthropic-risky-different-open-weight-rivals-china-2026-9 , model architecture, and attorney-client privileged information. Anthropic also said it holds the right to require METR to redact information. "METR is free to characterize our redactions publicly, and say whether it believes the redactions were justified, as long as it does so without revealing the redacted information," Anthropic said. The access does not extend to the wider METR team. The six chosen researchers, whose names were not immediately published on Thursday, will be "siloed" from the wider METR team. "Embedded evaluation https://www.machinebrief.com/glossary/evaluation gives a few qualified people working under Anthropic's security standards a close view of how we operate," Anthropic said. "It's not a way for other organizations to obtain our information, or that of our customers and partners." METR did not immediately respond to a request for comment on Thursday. Business Insider https://www.businessinsider.com/anthropic-ai-safety-watchdogs-metr-dario-amodei-2026-10 Get AI news in your inbox Daily digest of what matters in AI.