Anthropic announced that staff from Accenture, through its AI division Faculty, will work inside the company to evaluate and red-team models, with both firms committing at least $1 billion over five years. The choice of Accenture surprised observers who expected safety nonprofits like METR, Redwood Research, and Apollo Research; Anthropic says it is in talks with METR on pilot evaluations. Critics, per TechCrunch, view Dario Amodei's embedded-evaluator plan as a way to evade accountability, which Anthropic denies.
No score is assigned. Sources and their independence are shown in the citation chain below.