Anthropic announced that Accenture's AI division, Faculty, will embed evaluators inside the company to conduct red-teaming, alignment assessments, and safety testing, with over $1 billion invested over five years. The choice surprised many observers who expected AI safety research orgs like METR to fill the role. Anthropic plans to announce more evaluators soon and is in discussions with METR and nonprofits to pilot similar arrangements.
Background
Anthropic CEO Dario Amodei has been advocating for independent third-party evaluators embedded within AI labs to improve model safety and alignment. This initiative reflects growing industry attention to AI oversight and governance.
- Source
- TechCrunch
- Published
- Sep 19, 2026 at 05:44 AM
- Score
- 6.0 / 10