Anthropic has entered into a partnership with Accenture on the independent evaluation of frontier AI models as part of its efforts to strengthen oversight and safety practices.
The partnership comes amid a broader debate over independent evaluation of AI models, with questions also raised about Model Evaluation and Threat Research (METR), its independence and its ties to Anthropic.
The partnership will be led by Faculty, Accenture’s specialist AI business, and will include evaluating and red-teaming models, conducting alignment assessments and testing model safeguards, Anthropic said in a statement.
Anthropic and Accenture each expect to invest at least $1 billion in building capacity in this area over the next five years.
Anthropic is also in dialogue with nonprofit evaluator METR and other organizations on embedded evaluation.
Anthropic said the partnership is part of a commitment made by CEO Dario Amodei in his essay, “We Must Pace the Frontier,” to embed independent evaluators within the company.
In a September 13 post on X, David Sacks questioned whether METR is independent, citing what he described as connections between METR and Anthropic’s investors and staff.
“Stop pretending METR is independent when it is intertwined with Anthropic’s investors and staff,” Sacks wrote. He also questioned whether the same evaluators should assess AI companies that are not at the frontier.
Dario has written that we need to “pace the frontier,” and Sam has agreed. People may be surprised by my response: go ahead.
You guys are the frontier. By any reasonable metric — market share, revenue growth, model capability — the two of you have a duopoly on frontier…
— David Sacks (@DavidSacks) September 13, 2026
Five days later, Anthropic announced its partnership with Accenture, describing it as an important step toward the commitment outlined in Amodei’s essay.
According to Anthropic, unlike conventional external evaluators, embedded evaluators would work inside AI companies with access comparable to employees. This would allow them to observe models during training, follow decisions around their development and deployment, and communicate directly with staff.
The company said this access could help evaluators assess how an AI company operates, verify its safety commitments, identify blind spots and report incidents.
“Independent embedded evaluators do not reduce our accountability, but help to make it more verifiable. The safety of our models remains our responsibility,” Anthropic said.
Anthropic will fund Accenture’s work directly. It is also in dialogue with METR and other nonprofit evaluators to pilot elements of embedded evaluation using their own funding.
Anthropic said it expects frontier AI companies to work with several evaluators at once. The partnership with Accenture is non-exclusive, and Anthropic will work with other evaluators to be announced in the coming weeks, while Accenture will work with other AI developers in similar capacities.
Anthropic said there are currently no established standards governing what access embedded evaluators should receive or how they should report their findings. Nor is there a settled funding model for independent evaluation.
“Long-term, we think funding should come from pooled or government sources,” Anthropic said, adding that because neither system exists today, it plans to work with different evaluators under different funding arrangements.
Also Read: David Sacks Upholds Zuckerberg’s Approach: Don’t Call on Others, AI Safety is Routine Work
