Gary Marcus: AI extinction risk near zero, but catastrophe and dystopia risks loom
GaryMarcus · x · 2026-10-07
Sharing a roundup of his views, Gary Marcus puts the probability of AI-driven human extinction (P-Doom) in the next several years at "very close to zero," but argues the debate is obsessing over the wrong nightmare.
- P-Catastrophe: events capable of mass casualties—an AI hallucination contaminating an intelligence report, militaries acting on false information, automated systems disrupting power grids, malicious actors using cheap AI agents to manufacture propaganda or trigger crises
- P-Dystopia: gradual destruction of privacy, reliable information, and democratic institutions
His takeaway: "Don't panic about P-Doom, but worry a lot about P-Catastrophe and P-Dystopia."
More from AGI Musings
- Mathematician: norms for LLM use in serious math are still unsettled — littmath · 2026-10-07
- Jensen Huang emerges as AI doom's biggest foil, calls Altman and Amodei 'irresponsible' — AlexTensor · 2026-10-07
- You don't need 'unlearning', you need 1-2 years of financial runway — alexeyguzey · 2026-10-07
- China may be playing a different AI game: distribution over frontier model supremacy — ingliguori · 2026-10-07
- AI folks confuse having an idea with creating something, say artists and physicists — CatAstro_Piyush · 2026-10-07
- Neuroscientists debate whether brain properties left out of computational models matter for consciousness — eschwitz · 2026-10-07