The AI Regulatory Gap: Who Gets Frontier Models

joshgans · x · 2026-07-14

Responding to criticisms that a recent AI open letter lacked specificity, the author argues the real issue is that regulatory economics has almost no literature on constraining AI risks through concrete rules. They point out that current discussions either hover at the macro level of "should we pause AI" or offer broad governance principles. What's truly missing is detailed research on who should get access to frontier models, why, and how to design those thresholds.

Original post →

More from Safety

Safety channel →