DeepMind London Staff Vote to Unionise Over Military AI Contracts

Staff at Google DeepMind in London have voted to unionise with the Communication Workers Union and Unite the Union. The primary objective of this industrial action is to prevent the company from supplying its artificial intelligence models to the US and Israeli militaries. This move follows Alphabet's removal of a pledge against using AI for weapons development and surveillance from its ethics guidelines in February 2025.

The unionisation vote received 98 per cent support among Communication Workers Union members at DeepMind. Employees have submitted a formal letter to Debbie Weinstein, Google's managing director for the UK and Ireland, requesting joint representation. Specific demands include a commitment not to pursue weapons development, transparency over AI usage, and assurances regarding potential layoffs caused by automation.

The push to unionise began after the removal of the ethical pledge, which staff say signals a shift toward the further militarisation of the AI models being developed. In late February, staff at DeepMind and OpenAI signed an open letter supporting Anthropic following scrutiny from the US Department of Defense. This occurred after the Pentagon confirmed deals with seven leading AI companies, including Google, to use their models on classified networks under a clause allowing use for any lawful government purpose.

In 2024, Google fired more than 50 staffers in response to a protest over the company's military ties to the Israeli government, contributing to ongoing tensions. Recently, the US Department of Defense confirmed deals with seven leading AI companies, including Google, to use their models on classified networks under a clause allowing use for any lawful government purpose. In 2021, US-based Google employees formed the Alphabet Workers Union, which is not officially recognized by Alphabet for collective bargaining but has negotiated agreements for contractors.

If negotiations fail, the workers intend to seek arbitration to compel union recognition. The outcome of the ten-day union recognition period remains pending, with no official response from Google yet received. The extent to which potential research strikes would impact specific AI services like Gemini is currently speculative, though the campaign aims to halt the deployment of technology violating staff ethical standards.

The Communication Workers Union hopes this campaign will spur similar action at other frontier labs, such as those of Anthropic and OpenAI, which have recently expanded operations in London. John Chadfield, national officer for technology at the CWU, noted that workers at other frontier labs have seen what Google DeepMind workers have done and are asking for assistance.