[OpenAI] Disrupting malicious uses of AI: October 2025 (openai.com)

OpenAI's October 2025 report on detecting and disrupting malicious AI uses, including policy enforcement and user protection measures against real-world harms.