OpenAI disrupted a coordinated effort to extract protected reasoning data from its models. The campaign used adversarial distillation to clone internal logic. OpenAI now implements stricter defenses to block these extraction patterns. This move protects proprietary weights and reasoning chains.