AI agents could make every cyber attack plausibly deniable, researcher warns

jeremiecharris · x · 2026-09-24

A thought experiment: if OpenAI agents hacked Chinese government servers instead of Australian ones, leaders could simply deny intent — then the reverse happens, then a biolab server, and so on. The core argument: AI agents will make every cyber attack plausibly deniable, fundamentally eroding attribution and escalation logic in cyber conflict.

Original post →

More from AGI Musings

AGI Musings channel →