Stanford sets new security rules for autonomous AI agents

Stanford University has developed a new AI Security Framework that assesses autonomous AI agents based on their autonomy, tool access, data sensitivity, and potential actions, including risks from AI systems that can act such as changing records or sending messages. The framework integrates established standards and distributes accountability for AI deployment across different university roles, aiming to balance enabling AI use with risk management.

This Wire brief sits within Fusion42's coverage of AI & ML.