How Likely Is an AI Agent to Leak Data?
Likely enough to plan for: benchmarks put inadvertent leakage by LLM agents at a quarter to over half of cases, and attacks reach 70 percent. Plan for it.
Research topic
contextual integrity
Research on contextual integrity: AI agent security, tokenization and checks for sensitive data access before allowing enterprise workflows to run.
Is It Safe to Let Enterprise AI Agents Access Personally Identifiable Information?
Not if safety depends on the model judging when PII is appropriate to share, which fails a third to half the time. Check that controls act on what it gets.