DeepMind staff vote to unionize, citing opposition to AI military contracts
Key Points
- DeepMind staff voted 98% in favor of unionizing under the CWU and Unite the Union.
- Unions demand a clear commitment from Google to stop AI work that could be used for weapons or surveillance.
- Employees claim DeepMind's models already support military actions they consider illegal and immoral.
- Google has ten days to recognize the union before legal proceedings begin.
- Workers plan protests and a possible research strike targeting AI services like Gemini.
- The effort follows a broader employee push against Google’s Pentagon and Israeli‑military AI contracts.
- No comment from Google has been released as of now.
Employees at Google DeepMind's London headquarters have voted overwhelmingly to join a union, demanding that the company halt development of AI tools for military use. The Communication Workers Union and Unite the Union seek recognition as joint representatives for roughly 1,000 staff members. Workers argue that DeepMind's technology is already being used in conflicts involving Israel and the U.S., and they want guarantees that future projects will not violate ethical or legal standards. Google has ten days to respond before formal legal steps begin.
Staff at Google DeepMind’s London lab have taken a decisive step: 98 percent of Communication Workers Union members voted to unionize, and the union now seeks joint representation with Unite the Union for the lab’s roughly 1,000 employees. In a letter to Google management, the unions demanded that the company commit to stopping the development of weapons, surveillance tools, or any AI contracts that could harm people.
DeepMind workers say their models are already being used in ways they find morally unacceptable. An unnamed employee told the CWU that the technology is "complicit in violations of international law" and is "aiding Israel’s genocide of Palestinians." The employee added that even administrative uses, according to leadership, still make "genocide cheaper, faster, and more efficient." The staff want the right to refuse projects that conflict with their personal ethics.
Google now has ten working days to voluntarily recognize the union before a formal legal process forces recognition. If the bid succeeds, the union will give DeepMind staff a collective voice in negotiations over AI projects that could affect their roles, workloads, or job security.
Beyond the union vote, employees are planning additional tactics. Some are considering in‑person protests, while others are discussing a "research strike" that would pause work on improvements to Google AI services such as the Gemini assistant. The broader campaign targets Google’s military‑industrial AI contracts, which have drawn criticism from workers worldwide.
The move follows a recent open letter signed by hundreds of Google employees urging CEO Sundar Pichai to reject classified AI contracts with the Pentagon. Despite that pressure, Google, along with OpenAI and Nvidia, signed agreements allowing the U.S. Department of Defense to use their AI models for any lawful government purpose. Earlier this year, Google fired more than 50 staffers after a protest over the company’s ties to Israel’s military.
John Chadfield, national officer for tech workers at the CWU, framed the vote as a “really important moment” for tech labor. He said workers are linking their struggle to oppressed communities worldwide, grounding their demands in solidarity and trade‑unionism. The unions hope that collective bargaining will force Google to abandon contracts they deem ethically dangerous.
Google has not yet commented on the unionization effort. The outcome will test how far tech workers can push large corporations to align AI development with humanitarian standards.