Get reporting on AI risk, security and governance in your inbox.
The newsletter follows the same structure as the reporting: The signal, The risk, What researchers say and What I’m watching. Expect one substantial issue each week, with selective breaking posts when the evidence warrants them.
What you’ll get
- The signal: the development worth paying attention to.
- The risk: what could actually go wrong, and what evidence supports that concern.
- What researchers say: the strongest arguments and disagreements.
- What I’m watching: the next tests, papers, incidents and policy moves that could change the picture.
No hype cycle. No automatic catastrophe narrative. Just the evidence, the disagreement and what is changing.
AI Safety Watch 