AI
Inside OpenAI’s adversarial model-distillation incident
OpenAI says a coordinated campaign tried to extract protected reasoning. We separate the request counts, attribution, attack path, mitigations, and unresolved impact.
2026-10-04·Source OpenAI