Anthropic and Accenture Commit $1 Billion Over Five Years to Embedded AI Safety Evaluation

Anthropic partnered with Accenture's Faculty division to embed independent evaluators inside Anthropic with employee-level access to observe model training and deployment decisions, backed by a five-year, $1 billion-plus joint investment, with more non-exclusive evaluator partnerships planned alongside existing work with METR.