Turing Award Winners Caution Against Unsafe AI Deployments
Andrew Barto and Richard Sutton, winners of the 2025 Turing Award, have warned against the rapid and unsafe deployment of AI models, urging for thorough testing and proper engineering practices.
Published on March 29, 2025
Reinforcement learning pioneers Andrew Barto and Richard Sutton, honored with the 2025 Turing Award, have warned that the rapid deployment of AI models without sufficient testing poses significant risks. The researchers criticized tech companies for prioritizing quick business gains over rigorous engineering practices, a move that could lead to unintended and potentially dangerous outcomes.
Their groundbreaking work, which underpins major AI milestones including Google’s AlphaGo and OpenAI’s ChatGPT, continues to influence the field. Despite expressing concerns over safety and the impact of potential cuts in federal research funding on US leadership in technology, both researchers remain optimistic about the future benefits of reinforcement learning when applied responsibly. Their comments were publicly noted on March 5, 2025.