davidad on infohazards: don't suppress discussion of impending AI upheaval

davidad · x · 2026-09-05

Alignment researcher davidad lays out his bar for infohazards: information counts only if it makes others more capable of mass casualties, not merely if it's hard to predict how people react to the truth. Pushing back on the suggestion that talk of impending AI-driven upheaval could provoke despair or erratic behavior, he cites technocratic attempts to prevent pandemic panic — which biased the information environment and seeded distrust that still impedes public health years later. His threshold for censoring public discourse about big changes is far higher.

Related event: davidad: Don't hide AI upheaval to prevent panic(2 posts)→

Original post →

More from AGI Musings

AGI Musings channel →