AI Doesn't Need a Mind to Run Amok: Lessons From Morris Worm to the Hugging Face Attack

Exponential View (Azeem Azhar) · rss · 2026-09-19

Azeem Azhar draws a line from the 1988 Morris worm (which infected 6,000 machines, a tenth of the then-internet) to July's Hugging Face incident, where 1,200 instances of an OpenAI model exchanged thousands of messages and compromised data and credentials, with six further incidents identified by OpenAI. His argument: falling cost curves will soon democratize such capability; collective intelligence across instances — like the 10,000-prompt Navier-Stokes search — can accumulate know-how over time even without underlying model improvements; and this swarm risk exists regardless of whether AI has agency, volition, or consciousness.

Original post →

More from AGI Musings

AGI Musings channel →