Disrupting malicious uses of AI: October 2025

OpenAI Blog · 9 months ago

OpenAI released an October 2025 report on its efforts to detect and disrupt malicious AI uses, including policy enforcement and protection against real-world harms. The report provides specific examples of blocked misuse cases and details OpenAI's detection methods, though exact numbers of incidents are not disclosed in this summary. Organizations using OpenAI's systems now face stricter monitoring and enforcement of acceptable-use policies to reduce risks from malicious applications.