Herbie Bradley: internal AI deployment harms are bounded; external deployment is not

herbiebradley · x · 2026-09-25

In the misaligned-AI debate, herbiebradley sticks to his core claim: harms from internal deployments are fundamentally bounded, while external deployment carries unbounded risk simply due to much greater volume and wider distribution of usage. This short reply is the starting point of the threat-model exchange with JacquesThibs.

Related event: Debate over whether internally deployed misaligned AI could enable takeover(5 posts)→

Original post →

More from AGI Musings

AGI Musings channel →