Disrupting malicious uses of AI: June 2025

OpenAI Blog · 1 year ago

Anthropic released a report documenting methods for detecting and preventing malicious applications of AI systems. The report includes specific case studies demonstrating their detection and prevention techniques across different threat categories. The findings inform how Anthropic responds to identified misuse and guides their approach to safety measures in AI deployment.