LessWrong argues AI X-risk lacks viral power of political tribal fears
Political dread beats existential AI risk in memetic appeal, says new analysis.
The essay, published on LessWrong by Steff, explores why existential risk (X-risk) from AI fails to capture public attention compared to political tribal fears. The author references KatjaGrace's observation that the 'everyone dying at once' image feels like a thought experiment—abstract, neat, and not sinister—unlike concrete political threats. While fears like the Great Replacement theory or stolen elections go viral regardless of plausibility, X-risk strains credulity and feels too absurd.
The author proposes a provocative solution: deliberately politicize AI by having leaders like Sam Altman (OpenAI) and Dario Amodei (Anthropic) take extreme partisan stances. By planting rumors that each will use AI to steal the next election, public fear would skyrocket, creating demand for an immediate pause on AI development. The essay notes that nearly two-thirds of Republicans believe the 2020 election was stolen, showing how easily political narratives override evidence. The argument: if AI becomes a tribal issue, its risks will finally be taken seriously.
- Existential AI risk (X-risk) feels abstract and surreal, unlike visceral political fears like stolen elections.
- Nearly two-thirds of Republicans believe the 2020 election was stolen, despite no evidence.
- Proposal: Have AI leaders openly take extreme partisan stances to trigger viral fear and demand for an AI pause.
Why It Matters
This highlights how AI safety messaging may need to leverage partisan emotion to gain traction.