Analysis
Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle
Publication date unknown · Discovered research.google
Page published
Publication date unknown · First observed: 8 Oct 2026
Coverage timeline
Single-source analysis — one report is available.
Why it matters
Google's CAPS workshop report maps the foundational privacy and security gaps defenders must address as autonomous agents gain access to personal data and the ability to take consequential actions.
Google Research published a workshop report, "Open and Emergent Problems in Agentic Privacy and Security: A Contextual Angle," synthesizing input from 50+ academic and industry leaders at the Google CAPS Workshop. Framed around Contextual Integrity, the report outlines open research directions for securing autonomous AI agents at the system, model, user, and ecosystem levels, highlighting challenges like unstructured input ambiguity, adversarial prompt injection, and probabilistic control flows.