Anthropic and Accenture will embed Faculty evaluators inside Anthropic with access comparable to employees. The work will cover model evaluation, red teaming, alignment assessments and safeguard tests informed by enterprise deployments. Each company expects to invest at least $1 billion over five years. The operating model, publication rules and conflict safeguards are still being designed.
Anthropic and Accenture create an embedded frontier-evaluation model
2026-09-18Verified announcementWATCH
Anthropic and Accenture will embed Faculty evaluators inside Anthropic with access comparable to employees. The work will cover model evaluation, red teaming, alignment assessments and safeguard tests informed by enterprise deployments. Each company expects to invest at least $1 billion over five years. The operating model, publication rules and conflict safeguards are still being designed.
Imported from the existing research archive. Migration preserves the recorded judgment and is not a fresh verification.
What this means for you
For model buyers and safety teams, this is a concrete move from occasional external tests toward continuous third-party access inside a frontier lab. The arrangement is not yet proof of independent oversight: its value will depend on evaluator autonomy, what results become public and whether findings can block deployment.
Evidence & related material
Technical details & scores
| Model | embedded evaluators with employee-comparable access |
|---|---|
| Scope | model evaluation, red teaming, alignment assessment and safeguard testing |
| Evaluator | Faculty, Accenture's specialist AI business |
| Investment | at least $1B expected from each company over five years |
| Open Questions | operating model, publication rights, evaluator independence and deployment authority |
Technical impact 3
Engineering value 3
Adoption signal 4
Maturity 2
Verification cost 3
Risk 3