Accenture to test Anthropic’s AI safety amid growing concerns

Anthropic engaged Accenture and its Faculty unit for independent evaluation of frontier AI development and testing of safeguards. Both companies will invest at least $1 billion each in developing this area over five years.

• The evaluation will be conducted by Accenture’s Faculty unit with access to Anthropic workspaces
• External reviewers will be able to publish findings on risks without Anthropic editorial control
• Anthropic and Accenture will invest at least $1 billion each in this area over five years
• Anthropic is in talks with METR about a pilot for built-in evaluation at its own expense