What this is about
Identifying practical defenses and promising directions for future research.
This is a follow-up to SAGAI at IEEE S&P 2026 (Secure Agents for Generative Artificial Intelligence), but in a more intimate setting, where we hope to create personal connections and make space for deeper discussions.
We propose to run this workshop under the Chatham House Rule: participants are free to use the information received, but neither the identity nor the affiliation of the speaker(s), nor that of any other participant, may be revealed.
- 01
AI agents for security
How can we use AI agents to improve computer security? How can we enable them to investigate threats and respond to them?
- 02
Security for AI agents
How can we keep AI agents safe when they use tools, read outside content, and access sensitive systems? What limits should we put on their permissions and environment?
Further reading: Christodorescu et al. (2026). This paper was a product of the previous SAGAI workshops.
The program
One full day of technical exchange for a capped group of 60 invited participants, with approximately 15 graduate-student research posters.
- Invited technical presentations
- A graduate-student poster session
- Facilitated research breakouts
- Focused discussion across academia and industry
- Development of a post-event technical report
Who will be in the room
- Frontier AI laboratories
- Leading research universities
- UC San Diego faculty and research groups
- The systems, security, privacy, and AI research communities
- Graduate-student researchers
- Invited sponsors
Where we’ll meet
At Scripps Institution of Oceanography (the birthplace of UC San Diego), on the edge of stunning La Jolla and the Pacific.
Learn about the venue
