All incidents / Microsoft

Microsoft AI security incidents

1 security incidents involving Microsoft's AI since January 2025, including 0 where an agent acted on its own. The model was judged at fault in 1. Involvement does not mean Microsoft was responsible.