More time won't make policymakers decide better — an AI policy reality check

Afinetheorem · x · 2026-08-15

rSanti97 pushes back on a common assumption in AI policy: that policymakers make better decisions when given more time to act.

His counterexample is Social Security, which has been trending toward insolvency for 30+ years with policymakers fully aware — so it's neither an information problem nor a time problem, yet decisions haven't improved.

His takeaway: policymakers decide better only when a decision is required soon (but not too soon), they have the requisite information, and there are specific, useful levers to pull. The same author also shared a New Statecraft podcast episode on what CIA analysts actually do.

Original post →

More from Safety

Safety channel →