Research
Security, Privacy, and Agentic AI in a Regulatory View: From Definitions and Distinctions to Provisions and Reflections
University of Oslo researchers analyze 24 EU AI regulatory documents from 2024–2025, mapping where agentic AI's autonomous capabilities create regulatory gaps — particularly where the law is robust vs. porous on security and privacy provisions. Key finding: existing EU regulatory frameworks struggle to articulate precise stipulations for autonomous multi-step agent actions, leaving significant gaps that practitioners and compliance teams must account for. Useful for teams navigating EU AI Act compliance when building agentic systems.
↳ Follow the thread