Andrew Barto

Andrew Barto (born 1948) is professor emeritus at the University of Massachusetts Amherst, where he spent nearly five decades building the field of computational reinforcement learning. With his doctoral student Richard Sutton he developed the core algorithms of learning from reward and co-wrote the canonical textbook “Reinforcement Learning: An Introduction” (1998) — work that ultimately enabled systems like AlphaGo and the training methods behind modern chatbots. In March 2025 the pair received the 2024 ACM A.M. Turing Award, computing’s highest honor.

A laureate’s warning

Where many laureates use the Turing platform for a victory lap, Barto used his for an indictment. In a Financial Times interview after the announcement, he accused AI companies of releasing software to millions without safeguards, likening the practice to “building a bridge and testing it by having people use it.” He argued that sound engineering evolved precisely to mitigate technology’s harms — and that he saw no such discipline among the firms racing to ship generative AI, motivated as they are by monetizing giant data centers rather than by safety.

The criticism carried extra bite because of who shared his stage: co-laureate Sutton has argued humanity should welcome “succession” to digital intelligence, making the 2024 Turing Award perhaps the only one split between an AI-deployment critic and the field’s most cheerful successionist. Barto is no doomer prophesying extinction — his objection is professional, about process and responsibility, not science fiction.

Barto remains a builder at heart: he spent his career creating the very foundations today’s systems run on, and his complaint is that the industry deploying them has abandoned the engineering rigor the work deserves.