Research Scientist, Scalable Alignment (Google DeepMind) · Google · 2026
Research scientist on Google DeepMind's scalable alignment team who resigned after Google signed an agreement allowing the Pentagon to use its AI for classified work. Turner said the agreement lacked binding restrictions against autonomous weapons or mass surveillance, despite months of internal advocacy for enforceable safeguards. He wrote: “At that point, I couldn’t stay at Google in good conscience, so I left.”
Resigned in June 2026; the exact day is not publicly documented. Turner published his account of the departure on July 15, 2026.
“At that point, I couldn’t stay at Google in good conscience, so I left.”
Turner spent more than two years at Google DeepMind working on AI safety and scalable alignment. In his first-person account, he described trying to prevent Google from accepting an “any lawful use” classified military agreement without binding safeguards, drafting a proposed red-line and oversight framework, and escalating the proposal through senior leadership. After Google signed the agreement, Turner concluded that its language concerning autonomous weapons and mass surveillance was aspirational rather than binding and resigned without moving directly to another frontier AI laboratory.
Turner's first-person account of his resignation from Google DeepMind after Google signed an agreement allowing the Pentagon to use its AI for classified purposes. He describes months of internal advocacy, including a proposed red-line and oversight framework intended to preserve meaningful human authority over military AI and prohibit mass-surveillance uses. Turner argues that the agreement's restrictions were aspirational rather than binding and that Google retained no right to veto lawful government operational decisions. He says the decision made it impossible for him to remain at the company in good conscience.