Anthropic Picks Accenture to Help Put Its AI Slowdown Plan Into Action

TLDR

• Anthropic has selected Accenture as its first embedded evaluator for AI safety work.
• The companies will evaluate and red-team models, assess alignment and test safeguards.
• Anthropic and Accenture each expect to invest at least $1 billion over five years.
• The partnership is non-exclusive, with Anthropic planning to add more evaluators.
• The move follows CEO Dario Amodei’s proposal to slow AI development while stronger safeguards are built.

Anthropic has selected Accenture as its first embedded evaluator as the AI company moves forward with CEO Dario Amodei’s proposal to slow the pace of artificial intelligence development.

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 two companies will work together on model evaluations, red-team testing, alignment assessments and testing of safeguards.

Anthropic said the partnership is part of a broader effort to bring independent evaluators closer to the development process.

Anthropic Moves Ahead With AI Slowdown Proposal

Amodei published a three-step proposal on September 12 calling for a slower pace of AI development while stronger safety systems are put in place.

He argued that AI capabilities are improving quickly, partly because advanced systems can help build future generations of AI.

Amodei warned that this process could move faster than researchers and governments can understand or control.

The first step in his proposal called for independent evaluators to receive access similar to company employees.

Anthropic had already committed to that approach before announcing the Accenture partnership.

The company said Accenture and its AI business Faculty will work alongside Anthropic’s internal teams and existing safety partners.

Their role will include testing models for weaknesses, evaluating whether systems behave as intended and checking whether safety measures perform as expected.

Anthropic said the details of embedded evaluation are still being developed because the approach is relatively new.

Accenture and Anthropic Plan $2 Billion Investment

Anthropic and Accenture each expect to invest at least $1 billion over the next five years.

That would bring the combined expected investment to at least $2 billion.

Anthropic said there is not yet an established funding system for independent AI evaluation.

The company said it believes longer-term funding should eventually come from pooled industry or government sources.

For now, Anthropic plans to fund Accenture’s work directly because it views the safety work as urgent.

The partnership is non-exclusive.

Anthropic said it expects to announce additional embedded evaluators in the coming weeks.

Accenture will use expertise from Faculty, the AI company it acquired, as part of the work.

Faculty has experience testing and evaluating AI systems for leading laboratories and has worked on projects in government, defense, healthcare and infrastructure.

Accenture CEO Julie Sweet said embedded evaluation could become an important part of the AI safety process.

The announcement comes after Amodei’s slowdown proposal received mixed reactions from technology executives.

OpenAI CEO Sam Altman and Elon Musk responded positively to the idea, while Nvidia CEO Jensen Huang argued that slowing AI development through regulation was unnecessary.

Anthropic’s partnership with Accenture now puts the first part of Amodei’s proposal into practice, while the company prepares to add more independent evaluators.