IFP proposes 23 low-regret policies for the US to prepare for automated AI R&D

sjgadler · x · 2026-09-09

Former OpenAI researcher @aidanclark says he is asking for the first time whether things are moving too fast and calls for a clear definition of a successful pace. The reply points to an IFP report with 23 low-regret policy recommendations: transparency mandates on automated AI R&D risk management and incident reporting, at least $84M/year to resource CAISI with direct advisory power, clearer agency roles, intelligence coverage of foreign AI, and a risk strategy that accelerates defensive and commercial AI uses.

Related event: Ex-OpenAI Researcher First Questions Whether AI Is Moving Too Fast(3 posts)→

Original post →

More from AGI Musings

AGI Musings channel →