Hassabis Backs AI Pre-Release Safety Testing
Gary Marcus · rss · 2026-07-14
Gary Marcus reports that Google DeepMind CEO Demis Hassabis has publicly endorsed a "preflight safety testing" mechanism: frontier models should be submitted to standards bodies for review before large-scale deployment. Marcus mentions he has advocated for a similar system since 2023, arguing it should ideally be mandatory, transparent, and independent, akin to FDA reviews before drug releases.
In his article, Hassabis suggests that frontier labs could share models with standards bodies up to 30 days pre-release for review. If proven effective and robust, this protocol could be formalized, requiring frontier models to pass such tests before entering the US market. The article also stresses that labs should collaborate with standards bodies to fix critical post-release vulnerabilities. Marcus particularly appreciates the emphasis on independent technical experts.
More from AGI Musings
- Gary Marcus says LLMs still cannot really do math on their own — GaryMarcus · 2026-07-22
- Gary Marcus says LLM math skills are like knowing only a car’s engine size — GaryMarcus · 2026-07-22
- AI may make digital work infinitely leveraged while offline life gets more human — illscience · 2026-07-22
- Better AI math could save researchers time by killing false conjectures earlier — prateekj · 2026-07-22
- AI’s economic forecasts are split by nearly a quadrillion dollars by 2035 — bittingthembits · 2026-07-22
- Open source is becoming tech’s soft power, says Kevin Xu — kevinsxu · 2026-07-22