Tracks

Hands-on technical track on controlling and evaluating AI systems. The introductory module works through Redwood Research's "AI Control: Improving Safety Despite Intentional Subversion" paper and then rebuilds its trusted-monitoring result as an interactive, model-backed demo.

content coming soon

Interactive modules, primers, and exercises on verifying international AI agreements — from why the problem matters, to the compute supply chain, to the game theory that makes verification the hinge of any deal.

A demo track showing every feature in one place — the content types, each exercise type, an inline paper with embedded activities, an end-of-module assessment, and a cross-module prerequisite.