The article discusses warnings from prominent AI safety advocate Max Tegmark, who urges artificial intelligence companies to rigorously assess the existential risks posed by advanced AI systems. Drawing a parallel to safety calculations made before the first nuclear test, Tegmark proposes that firms calculate the probability—termed the “Compton constant”—that a superintelligent AI could escape human control. Doing so, he argues, could help drive global consensus and regulation for AI safety, especially as the race to create more powerful AI models intensifies. The piece also references the Singapore Consensus report, which outlines broad research priorities for AI safety, and highlights renewed momentum for international cooperation following recent political debates over the future of AI governance.
Related article:
Why Experts Are Worried About AI—and What They Say We Should Do About It
