If Anyone Builds It, Everyone Dies

Eliezer Yudkowsky and Nate Soares present a rigorous, high-stakes argument against the development of superintelligent artificial intelligence. Framing the issue as a variant of Pascal's Wager, the authors contend that the potential for existential catastrophe outweighs any conceivable benefits, urging a global moratorium on AI research. The text synthesizes technical insights on machine learning and alignment with philosophical inquiries into risk, morality, and the future of the species.

The writing style is urgent, analytical, and polemical, blending the precision of a technical white paper with the rhetorical force of a manifesto. It avoids speculative fiction tropes in favor of a direct, logical appeal to policymakers and the public. By treating the prospect of human extinction as a near-certainty if superintelligence is achieved, the book positions itself at the intersection of computer science, ethics, and civilizational strategy, demanding immediate regulatory action to prevent irreversible harm.

Description adapted from Open Library ↗.