Anthropic gives third-party auditors ’employee-like access’ in AI industry first

ZDNET’s key takeaways

• Anthropic announced it would give METR evaluators “employee-like access.”
• AI labs are under increased regulatory pressure following security incidents.
• Third-party testing is not yet required under law.

On Thursday, Anthropic announced that it will “give employee-like access” to six evaluators from METR, a third-party nonprofit that tests frontier models.

Also: Claude sent a false homicide tip to police – and it’s part of a troubling AI pattern

The employees have been given Anthropic hardware and office space, and can access “internal tools and systems that are roughly on par with what our own risk assessment teams have access to,” Anthropic said. “Outside a small number of clearly defined categories (e.g., data about customers, attorney-client privileged information, or information about our model architectures), Anthropic employees are encouraged to share information with them just as they would with a fellow employee working on alignment risk.”

METR will publish its findings, none of which Anthropic is allowed to withhold, but the company will ask METR to redact sensitive information where necessary.

Anthropic CEO Dario Amodei proposed embedding evaluators from nonprofits like METR in a September blog post, one of three suggestions for “pacing the frontier.”

In the announcement, Anthropic said that it wants to be accountable, but “can’t publicly share all the information relevant to our safety work,” because “some of it could be useful to malicious actors, or is commercially sensitive, like compute or training information.” That’s where METR will come in: Anthropic will give them “deep access” to models and company information, and METR will share its conclusions about the risks those models pose.

Also: Regulators are trying to protect you from being fired by AI – here’s how

“We aim to close the gap between what evaluators understand about a frontier AI company’s systems and what the company itself understands,” Anthropic added.

Mandating evaluations

As US AI regulation has remained unsatisfactory to many safety researchers and security advocates, testers have suggested mandating third-party review of models, either before release or otherwise as part of the development process.

The FRONTIER Act, a bipartisan bill introduced in July, would require certain AI companies to allow “independent verification organizations,” or IVOs, to review the safety of their development process. OpenAI expressed at least partial support for the bill in mid-September, the first time it had backed any proposed federal regulation of AI, despite at times being in favor of unified federal regulation over a patchwork of state AI laws.

Also: How many jobs is your state losing to AI?

Thus far, regulation at the federal level under the Trump administration has largely meant voluntary agreements. Major AI labs have signed several pledges with the US government to proactively share concerns about models and work with external auditors, but those pledges are not binding. That efectively leaves the AI industry to self-regulate.

In an exclusive statement to Axios, the Super Intelligence Force — named for the term a recent White House mandate wants to see as a replacement for “artificial intelligence” going forward — said AI companies now “must immediately disclose incidents involving their models and follow with swift, decisive action to remedy any and all harm.” It’s unclear what about this incident prompted the requirement, as it follows a summer of high-profile breaches by frontier lab models during testing and was relatively less destructive.

Also: Who’s responsible for catching rogue AI agents? You are

Without knowing what lobbying efforts preceded Anthropic’s announcement and the government’s change of heart, it’s hard to say whether Anthropic is simply proceeding with its own existing plan or getting ahead of the FRONTIER Act passing in the future.

Radhika Rajkumar

Senior Editor

Radhika Rajkumar is a senior editor at ZDNET based in New York City. She covers AI, specializing in safety, privacy and security, policy, education, and synthetic media. She also leads ZDNET's newsletter strategy. Radhika holds a Masters in Creative Publishing and Critical Journalism from The New School. See full bio