Anthropic has announced a collaboration with consulting firm Accenture, aiming to invest at least $2 billion over the next five years in the independent assessment of advanced AI models.
We’re partnering with Accenture on independent evaluation of frontier AI—part of our recent commitment to embed evaluators at Anthropic. Both we and Accenture expect to invest at least $1 billion to build capacity in this area over the next five years. https://t.co/SHVzjpgnfx
— Anthropic (@AnthropicAI) September 18, 2026
The initiative will be led by Faculty, a specialized AI division of Accenture. Plans include evaluating models, fine-tuning, alignment checks, and testing safety mechanisms.
This initiative aligns with the concept of "embedded evaluators," which Anthropic CEO Dario Amodei discussed in June. According to the company, these teams should operate within AI developers with access comparable to that of employees. This arrangement would enable monitoring of model training, deployment, internal decisions, engagement with staff, verification of safety commitments, and identification of vulnerabilities in processes.
In September, Amodei called for a slowdown in the development of advanced AI models due to increasing security risks. He cited the incident involving OpenAI and Hugging Face as one of the arguments for this caution. Amodei proposed three approaches to mitigate risks: international agreements, unified industry standards, and independent audits within laboratories.
The partnership with Accenture is not exclusive; Anthropic plans to announce collaborations with other evaluators in the coming weeks, while Accenture will also work with third-party AI developers in a similar manner. Concurrently, discussions are ongoing with the nonprofit organization METR regarding a pilot for embedded evaluation funded by them.
In the long term, Anthropic envisions funding for independent evaluations to come from pooled or governmental sources, as outlined in the Advanced AI Framework initiative. Currently, there are no such mechanisms in place, and the company is directly financing Accenture's work, acknowledging that this creates a clear conflict regarding the evaluator's independence.
It is worth noting that Nvidia CEO Jensen Huang has opposed regulatory slowdowns in AI, labeling the safety of the technology as an "engineering problem."