OpenAI Says It Disrupted Coordinated Model Distillation Campaign

OpenAI said it identified and disrupted a coordinated campaign aimed at extracting protected reasoning from its models. The company said the activity began in early July and involved attempts to use model outputs in ways that could help reproduce or improve another model without authorization. According to OpenAI, the operators did not breach encryption, access databases, or obtain direct access to stored user conversations. Instead, they manipulated model interactions to surface hidden reasoning in visible form. OpenAI said the issue reflects a broader security challenge for advanced AI systems and shared details with industry partners to help strengthen defenses. OpenAI said it responded with account bans and restrictions, stronger signup and infrastructure controls, and expanded monitoring. It also said it added protections for hidden reasoning, closed a replay path that could expose encrypted reasoning, and worked with thirdparty providers when activity moved through outside services. The company said it will continue improving technical safeguards and threatinformation sharing.