#multi-agent-systems

Google Research Proposes Contextual Integrity Policy Engines and Multi-Agent Benchmarks for Agentic Privacy

A Google Research workshop report proposes LLM-driven policy engines based on contextual integrity, plus standardized multi-agent benchmarks, to make AI agents more privacy-aware, auditable and safer to deploy. What it means for data and AI engineers.