Accenture up 6% on Anthropic AI safety evaluation partnership

Anthropic CEO Dario Amodei has called for better transparency in the AI sector, highlighting the need for third-party evaluators to ensure companies adhere to training and deployment guardrails. To this end, Accenture has partnered with Anthropic to independently evaluate its frontier AI developments. This collaboration, led by Accenture's Faculty, will involve conducting assessments and testing model safeguards at Anthropic.

Amodei emphasized that such embedded evaluators, equipped with permissions and tools similar to internal employees, can guarantee transparency and provide an unbiased, commercial-free opinion on AI safety practices. Both companies plan to invest at least $1 billion over the next five years to build capacity in this area, with Anthropic funding the evaluation work directly due to its urgency.

Though this partnership marks a significant shift, Anthropic is not alone in advocating for cautious AI development, as Sam Altman has also called for an OpenAI public listing amid safety concerns.