July 15, 2026
AI, angst, and a dramatic exit
I Left Google DeepMind
After warning Google about military AI, one insider walked — and the comments exploded
TLDR: A Google DeepMind insider says he quit after failing to stop the company from making military AI deals without strong safety limits. Commenters were split between praising a rare moral stand and roasting it as naive, with the loudest message being: big companies care about contracts, not internal protests.
A former Google DeepMind researcher says he tried for months to stop the company from selling powerful artificial intelligence tools to government agencies and the military without hard limits on deadly weapons or mass spying — and when those efforts went nowhere, he quit. But online, the real fireworks were in the replies, where readers split into Team "respect the conscience move" and Team "welcome to reality, buddy."
One camp treated the resignation like a rare act of moral backbone in Big Tech. One commenter said they had "nothing but respect" for someone willing to take a stand when everyone else stayed quiet, arguing that more people need to risk their careers over the choices huge companies make. The other camp was much harsher — and much snarkier. The biggest dunk was basically, "scientist discovers reality," with critics saying giant corporations were never going to put ethics above contracts, power, and profit.
Then came the spicier fight: was this principled, or just impractical? One commenter argued that if the line is "a government killed innocent people," then good luck finding any country to do business with. Another mocked the idea that rival AI firms would unite for ethics, comparing it to bitter competitors teaming up over app store rules. And hovering over all of it was one brutally cold piece of advice: if you hate what Google is doing, leave — they do not care about you. He did. The comments still can’t decide if that makes him brave, naive, or both.
Key Points
- •The article says the author left Google DeepMind after Google signed a military AI deal the author believed lacked restrictions against autonomous weapons and mass surveillance.
- •The author links their activism to learning that Google sells cloud services to agencies within the Department of Homeland Security.
- •The essay states that the author sought support from senior AI figures, including Stuart Russell and Jeff Dean, but says those efforts did not produce the desired intervention.
- •The author says they wrote a 25-page proposal with contract language and oversight mechanisms and sent it to Demis Hassabis, but it was not adopted.
- •The article cites the DHS 2025 AI Use Case Inventory as listing Google among generative AI providers used to improve DHS operational efficiency.