The case that AI existential-risk warnings may be counterproductive

In a new essay, the authors challenge two narratives in AI safety debates: that existential risk is imminent, or that warnings about it are insincere hype or regulatory capture. They argue the warnings could instead be sincere but wrong, and that an existential-risk framing may be counterproductive for AI safety policy. They also argue that we systematically underinvest in resilience against catastrophic and systemic risks.