Anthropic and Accenture plan to build a team of independent evaluators who will work inside Anthropic as its frontier AI models are trained and deployed. Each company expects to invest at least $1 billion over the next five years, putting the combined plan at a minimum of $2 billion.
The partnership, announced by Anthropic and Accenture on Friday, will be led by Faculty, Accenture’s specialist AI business. Faculty’s team will evaluate and red-team models, conduct alignment assessments, and test the safeguards used around them. Reuters reported the agreement as AI developers face pressure from regulators, companies, and researchers to show that advanced models are safe and reliable.
What embedded evaluation means
Anthropic says the evaluators will have access comparable to an employee’s. They will be able to watch models during training, follow decisions about how systems are built and deployed, and speak directly with staff. That access gives evaluators a view of how safety decisions are made inside the company, rather than limiting the review to a finished model.
The evaluators will also be able to report incidents and give the public a more informed account of the benefits and risks they observe. Anthropic says the arrangement will not shift responsibility for model safety away from the company. Its models remain Anthropic’s responsibility, while the evaluators are meant to make the company’s safety commitments easier to verify.
Funding and standards are still unsettled
The system is still being designed. Anthropic says there are no standards yet for the information evaluators should receive or how they should report their findings. There is also no settled system for funding independent evaluation. For now, Anthropic will fund Accenture’s work directly. The company says long-term funding should come from pooled or government sources, but neither system exists today.
The announcement sets up the work with Accenture while those standards and funding arrangements are still being developed. Anthropic has not yet defined a common framework for access to internal work or for reporting findings.
Anthropic plans to add more evaluators
The partnership is non-exclusive. Anthropic says it is in discussions with METR and other nonprofit evaluators about piloting parts of the same model with their own funding. It expects to announce additional evaluators in the coming weeks, while Accenture plans to work with other AI developers in similar roles.
Anthropic says frontier AI needs an ecosystem of evaluators operating with shared standards. The company has not yet set those standards, so the Accenture partnership is an early implementation of an approach that is still being developed.
Sooraj Raj is a programmer, blogger and professional developer with an interest in technology research and publishing. He writes about technology developments and shares information in a clear, straightforward way.
