Anthropic moves fast on AI safety concerns with Accenture “embedded evaluator” partnership

Anthropic has announced a partnership with Accenture to improve AI development processes amidst rising safety concerns.

Under the deal, Accenture will act as an “embedded evaluator”, conducting assessments of frontier AI models. Accenture’s specialist AI business, Faculty, will lead on the project.

In a blog post detailing the move, Anthropic revealed this will include red-teaming of AI models, alignment assessments, and testing of model safeguards.

“Accenture helps businesses and governments deploy AI across many industries,” Anthropic said. “Their understanding of how enterprises use AI in practice informs their safety approach, and they will bring that perspective to evaluating our models.”

Both organizations expect to invest at least $1 billion to expand capacity in this area over the next five years.

How Accenture’s embedded evaluators will work

According to Anthropic, embedded evaluation of AI models is a new frontier within AI research, and exact details on how this will pan out are “still being worked out”.

What we do know is that the deal will see Accenture AI specialists embedded directly within companies, rather than acting as external evaluators. Anthropic said embedded specialists will be given access “comparable to an employee’s” when conducting tests.

“That access allows them to watch models take shape in training, follow the decisions that govern how those models are built and deployed, and speak directly to employees,” the firm explained in a blog post.

With intimate access to training and development processes, Anthropic said this will allow evaluators to “assess how a company operates, verify that it is keeping its safety commitments, and identify blind spots”.

These specialists can also report incidents and, ultimately, provide the public with a “more informed account” of the potential benefits and risks associated with new frontier models.

Long-term, Anthropic said it expects frontier labs to work with multiple evaluators at once, noting that the partnership with Accenture is non-exclusive.

“Anthropic will work with other evaluators to be announced in the coming weeks, and Accenture will work with other AI developers in similar capacities,” the company said.

Rising AI safety concerns

The move by Anthropic comes amidst growing concerns over AI-related safety and security risks. Earlier this month, CEO Dario Amodei published a lengthy essay urging for a slowdown in AI development to counter potential harm.

Amodei’s call to action was backed by leading industry figures, including OpenAI CEO Sam Altman and xAI chief Elon Musk.

As ITPro reported last week, however, not all industry leaders are on board with calls to “pace the frontier”.

During an appearance at Dreamforce 2026, Nvidia CEO Jensen Huang took a bullish stance on the pace of development, suggesting that responsibility for safety and security rests firmly on the shoulders of individual companies.

“If we're not confident about the safety of the products, like all companies … if you build a product or a service and you're not confident in its functionality, capability, or safety, then don't release it,” he told Salesforce CEO Marc Benioff.

FOLLOW US ON SOCIAL MEDIA

Follow ITPro on Google News and add us as a preferred source to keep tabs on all our latest news, analysis, views, and reviews.

You can also follow ITPro on LinkedIn, X, Facebook, and BlueSky.