Safety
Are AIs about to subjugate humanity? In the debate about catastrophic AI risk, evolutionary scenarios receive far too little attention compa…
Are AIs about to subjugate humanity? In the debate about catastrophic AI risk, evolutionary scenarios receive far too little attention compared to largely speculative arguments about “instrumental con
Are AIs about to subjugate humanity? In the debate about catastrophic AI risk, evolutionary scenarios receive far too little attention compared to largely speculative arguments about “instrumental convergence” (aka “you can’t fetch the coffee if you’re dead”). Yet evolution is a known mechanism for producing selfish, power-seeking agents—and it is substrate-neutral, operating in silico just as readily as in carbon. That’s why it’s encouraging to see three heavyweights in evolutionary biology and AI weigh in with an important new paper in PNAS, building on arguments from @hendrycks and my own work with @simonfriederich. I'll write a post about the paper soon! https://www.pnas.org/doi/10.1073/pnas.2527700123?__readwiseLocation=
Source: Dan Hendrycks (X) | 2026-05-02