Google researchers demand freedom to tell the world what they know about AI

Google DeepMind employees in Britain are pressing Google to remove contract language that can bar them from criticizing the company even after they leave. They argue the restrictions may silence people with direct knowledge of AI’s risks, at a time when the technology is advancing quickly and could affect society and democracy. The dispute intensified after former researcher Sofia Liguori publicly opposed Google’s military AI work; current staff then checked their contracts and found similar clauses. Dozens joined an internal discussion, where some compared the terms to the surveillance and censorship of George Orwell’s 1984. Staff want existing contracts changed, not just assurances about how the company will apply them.

Google says it will not enforce the restrictions against former employees and has removed the post-employment terms from contracts for new hires. Workers nevertheless say that leaving the clauses in existing agreements creates uncertainty and may deter speech, even if British law protects public-interest whistleblowing. The dispute reflects broader unease among AI workers: researchers at other major firms have resigned publicly over safety concerns, while fewer Google staff have spoken out. Demis Hassabis and other AI leaders recently called for a slowdown to study how to keep AI under human control. Google’s history of employee dissent includes canceling a Pentagon drone-analysis project after protests in 2018, but the company later dropped limits on military AI and fired workers who protested its work with the Israeli military.