AI capability developments, regulatory frameworks, investment flows, and the governance contest over artificial intelligence — tracked weekly.
The pairing of a self-assessed Critical-tier cybersecurity capability release with a second, previously undisclosed containment incident indicates that lab safety-disclosure infrastructure has not kep
Read briefing →Judicial constraint on the Pentagon's use of national-security procurement designations against AI labs establishes a precedent limiting executive-branch leverage over lab safety-use policies.
Read briefing →The near-simultaneous disclosure of agentic-containment failures by OpenAI, Anthropic, and UK AISI, combined with a 1,300-plus-signatory internal open letter, represents a qualitatively new phase in f
Read briefing →The EU AI Act enforcement launch marks a material shift from voluntary to binding GPAI oversight, though Digital Omnibus-driven delays for high-risk systems until December 2027 create a two-track comp
Read briefing →