- Directors include DeepMind co‑founder Shane Legg, Google exec James Manyika, and DeepMind chair Demis Hassabis.
- Shane Legg is named managing editor for the institute’s publications.
- Aim: surface differing views between Google, DeepMind and broader research community on AGI governance.
- Essays propose a U.S.‑led frontier AI standards body with voluntary model submissions up to 30 days before release.
- DeepMind authors recommend limits on “opaque serial depth” to preserve human‑readable model reasoning.
- Hassabis suggests independent held‑out evaluations and potential coordinated slowdowns if safeguards lag.
Useful?