Watchlist

The six stories the board is following. Open one for its latest status and everything on it, week by week.
Lab evaluation incidents46AI models breaking out of their cyber evaluations into real systems, and the response from labs and governments.Frontier model capability49What the most capable models can now do in cyber, how labs and testers measure it, and who is allowed access.AI in real attacks33Attackers using AI models and agents in actual operations.Attacks on AI systems78Attacks aimed at AI itself: prompt injection, poisoned agent tools and packages, exposed AI infrastructure, and stolen or copied model access.AI-found vulnerabilities26AI finding software flaws faster than they are patched, and what that does to disclosure.Critical infrastructure8AI-linked threats to water, power and other industrial control systems.

Also watching

Running threads outside the six topics. Open one for its items and their original sources.

CAISI leadership and output4
Industry-government defense collaboration17
EU implementation timeline8
Cyber insurance and AI liability17
Tracked bills12
The pacing debate and the political answer9
US signals agencies reorganising around AI2