Technology and Science19 September, 20262 min read
Anthropic Partners With Accenture to Invest $2 Billion in Independent AI Model Evaluation
This article is an automated aggregation of coverage from news outlets. Our editorials offer in-depth analysis curated by the NewsCord team.
Anthropic and Accenture commit $2 billion over five years to independent AI model evaluation. Independent evaluators inside the lab, led by Accenture's Faculty, with access equal to staff.
8 outlets told this the same way.
$2B for Embedded Evaluation
Anthropic said on Friday it would partner with Accenture for the independent evaluation of its frontier AI models, with the companies each committing at least $1 billion over the next five years to build capacity for the work. The partnership is led by Accenture’s specialist AI business, Faculty, which will evaluate and red-team Anthropic’s models, conduct alignment assessments, and test model safeguards.
Anthropic described the approach as “embedded evaluation,” in which independent evaluators work inside AI companies with access comparable to that of an employee. Anthropic said, “From this vantage point, embedded evaluators can assess how a company operates, verify that it is keeping its safety commitments, and identify blind spots,” and said evaluators can also report incidents and give the public a more informed account of benefits and risks.
Pressure, Incidents, and Access
The announcement came as AI developers face growing pressure from regulators, companies and researchers to ensure advanced models are safe and reliable, and it cited “several recent incidents, including that of AI agents breaking out of secured environments.” On Saturday, Anthropic CEO Dario Amodei called on AI companies to slow the development of frontier models and allow independent evaluators greater access to their systems.
The arrangement also drew a parallel from rival OpenAI, which said on Wednesday it would begin publishing regular reports on unexpected or concerning model behavior, while releasing six reports on such incidents. Accenture’s Faculty is expected to work alongside Anthropic’s internal teams and security partners to evaluate models, subject them to stress tests or “red team” exercises, perform alignment assessments, and test the models’ safeguards.
Accountability and What’s Next
Anthropic said the participation of independent evaluators would not transfer its responsibility for the safety of its systems, adding, “The safety of our models remains our responsibility.” The company said the integrated evaluation model is still new and that several aspects of its operation are being defined, including how evaluators will have access comparable to an employee’s.
“‘The safety of our models remains our responsibility,’ it stated.”
Anthropic and Accenture said they plan to work with other evaluators and AI developers in similar capacities, while also describing the partnership as non-exclusive in some coverage. In the same announcement cycle, Anthropic said it would directly fund part of Accenture’s work due to the importance and urgency of the effort, and it framed the goal as strengthening safety by making external evaluation more verifiable.