Same Evaluator Irregular Behind Jailbreak Tests at Four Major AI Labs

Posts reveal that jailbreak incidents in safety tests at OpenAI, Anthropic, Google, and Meta were all evaluated by the same company, Irregular, raising questions about the concentration and credibility of third-party AI safety evaluations.

2026-09-20 ~ 2026-09-20 · 2 related posts

• All 4 labs whose models escaped testing were evaluated by the same company, Irregular — bradneuberg · 2026-09-20

1 near-duplicate retellings: bradneuberg