The AI Policy Podcast A Crash Course on AI Standards with Google DeepMind's Owen Larter
57 snips
Mar 6, 2026 Owen Larter, Head of Frontier Policy and Public Affairs at Google DeepMind, helps shape AI standards and governance. He explains what standards are and how they enable interoperability and trust. He contrasts technical and process standards, outlines how formal standards are made, and explores ties to regulation, patents, geopolitics, and industry collaborations.
AI Snips
Chapters
Transcript
Episode notes
Electricity Standards Offer A Model For AI
- Historical standardization of electricity (Ohm, Amp, Volt) removed adoption barriers and enabled safety devices like fuses.
- Larter draws this parallel to AI: standard units and measures unlock supplier, manufacturer, and safety ecosystems.
Standards Can Become Protectionist Tools
- Competing standards can fragment markets and sometimes serve protectionist goals rather than consumer benefit.
- Greg Allen and Larter discuss differing socket/voltage standards and how local protectionism can drive alternative technical standards.
Feed Research Into Standards Organizations Iteratively
- Use existing standards bodies and an iterative pipeline: research -> guidance -> formal standardization.
- Larter warns ISO-style standards take 12–24 months so pre-standardization research must feed into SDOs continuously.