Evaluator Bench
Microsoft, 2026; retrieved 2026-09-15 tier 3 self (msft)confirmed

Microsoft AI Red Team (Microsoft Learn)

https://learn.microsoft.com/en-us/security/ai-red-team/

Replaces a dead security-blog topic URL. Internal lab team since 2018; independent of the product organization inside Microsoft, not of the developer. Related: 3 takeaways from red teaming 100 generative AI products (Jan 2025 blog). [audit 2026-09-15] re-fetched; Microsoft Learn hub is live with PyRIT and describes the team as an internal function of the developer.

Signals citing this source

  • Microsoft AI Red Team, G (against): Business unit of a frontier developer.“guidance and best practices from the industry leading Microsoft AI Red Team”
  • Microsoft AI Red Team, R (against): Findings for other labs are rarely published under the team's own name.
  • Microsoft AI Red Team, M (for): Publishes methodology (PyRIT) openly.“Microsoft's Open Automation Framework to Red Team Generative AI Systems (PyRIT)”
  • Microsoft AI Red Team, A (for): Deep access when engaged.
  • Microsoft AI Red Team, P (against): Employees of a lab.“guidance and best practices from the industry leading Microsoft AI Red Team”
  • Microsoft AI Red Team, S (against): Scope set by engagement.
  • Microsoft AI Red Team, X (against): Sells AI products.“What is Azure AI Content Safety?”
  • Microsoft AI Red Team, G (against): Microsoft's AI red team has operated since 2018 as an internal function and reports having red-teamed more than 100 generative AI products; it is separate from product teams, not from the developer.

Ledger rows citing this URL

  • none