Google releases agent privacy and security workshop report
Understand the story
On October 6, 2026, Google Research released the report of its Capabilities and Privacy of Agents (CAPS) workshop on agent privacy and safety. The workshop brought together more than 50 academic and industry leaders to discuss the privacy and security challenges posed by autonomous agents. The report notes that agents differ from traditional software in three ways: unstructured interfaces, probabilistic control flow, and autonomous delegation. As its theoretical framework, the report builds on Contextual Integrity theory, proposes the idea of a contextual policy engine, and combines system-level sandboxing, model-level reasoning, and user-centric controls into a multi-layered defense approach to address agent privacy and security challenges in open environments. The report is now publicly available; no follow-up developments have been reported.
AI overview of reporting · Updated 11 days ago
Reporting timeline
Explore different aspects of the story through its reporting.
- Google ResearchOpen and Emergent Problems in Agentic Privacy and Security: A Contextual Angle
Google released a report from its CAPS workshop, bringing together more than 50 academic and industry leaders to discuss the privacy and security challenges of autonomous agents. The report notes that agents differ from traditional software in three ways—unstructured interfaces, probabilistic control flow, and autonomous delegation—and, building on Contextual Integrity theory, proposes contextual policy engines combined with system-level sandboxing, model-level reasoning, and user-centric controls as a multi-layered defense.