Google DeepMind’s first formal talks with London employees seeking union recognition have already produced a procedural fight, according to WIRED, with union representatives saying the company failed to engage seriously and Google saying the process is moving as expected.
The dispute matters because the workers are trying to force a collective voice inside one of the world’s most influential AI labs, after Alphabet changed its AI ethics guidance and Google expanded government AI work. The immediate issue is narrower and more bureaucratic: who gets represented, who sits in the room, and whether Google will voluntarily recognize the unions.
In May, DeepMind employees asked Google to recognize the Communication Workers Union and Unite the Union as joint representatives, WIRED reported. Google declined that request, but agreed to take part in talks overseen by an outside arbitrator.
The first meeting took place Wednesday and included union officers, DeepMind employees involved in the organizing effort, the arbitrator, and DeepMind human resources representatives, according to WIRED. John Chadfield, a CWU officer who attended, said the absence of senior DeepMind management signaled that the company was not approaching the talks in good faith.
“Recognition talks not being attended by senior management at the opening stage is a leading indicator that a company isn’t engaging in good faith,” Chadfield told WIRED. He said negotiations had stalled early.
Google DeepMind rejected that characterization. Al Verney, a company spokesperson, told WIRED the first task was to define the group of workers the unions want to represent, and said the parties agreed on next steps for that work. Verney said “appropriate representatives” attended the meeting.
During the session, a DeepMind employee read a prepared statement on behalf of pro-union colleagues, according to WIRED, which reviewed the letter. The statement accused Google DeepMind of routing workers’ concerns through HR rather than holding substantive discussions with them.
The letter also alleged that Google had limited internal discussion about the union effort by shutting down or altering internal chat spaces and blocking staff from replying to company-wide communications about the campaign. It claimed employees who tried to work around those limits were reprimanded by HR. Multiple people familiar with the meeting told WIRED that HR representatives interrupted the employee reading the statement twice.
An anonymous DeepMind employee involved in drafting the letter told WIRED the restrictions were intended to intimidate workers and described them as union-busting tactics. Verney said Google DeepMind would keep engaging in the process and said employees still have “a variety of other channels and opportunities” to discuss views on other topics.
The organizing push began in February 2025 after Alphabet removed language from its AI principles saying it would not use AI for weapons development or surveillance, WIRED previously reported. One anonymous DeepMind employee told WIRED those principles were a major reason they joined the lab.
AI workers have been contesting military uses of their systems across several companies. WIRED reported that DeepMind and OpenAI staff signed a February open letter supporting Anthropic after the US Department of Defense sought to label the company a supply chain risk over limits on autonomous weapons and mass surveillance uses. The New York Times later reported that Google had agreed to let the Pentagon use its AI for “any lawful government purpose,” and The Washington Post reported that about 600 US-based Google workers signed a protest letter.
The Defense Department later said it had struck deals with seven AI companies, including Google, SpaceX, OpenAI, and Microsoft, for use of their models on classified networks. Google spokesperson Jenn Crider told The New York Times in April that the company was proud to support national security and remained committed to limits on domestic mass surveillance and autonomous weaponry without appropriate human oversight.
If the London talks do not advance, Chadfield told WIRED the workers would ask the UK’s Central Arbitration Committee to compel Google to recognize the unions.
This story draws on original reporting from WIRED.
