DeepMind union talks hit early snag over HR-only attendance

Key Takeaways

• DeepMind's first union negotiation meeting was attended only by HR, frustrating union representatives who called it a 'time-wasting exercise'
• The unionization push began after Google removed ethics pledges about weapons and surveillance from its AI guidelines
• If talks fail, UK unions can petition an arbitration committee to force Google to recognize them

Google DeepMind's first formal unionization meeting ended with workers and union officials frustrated and calling the session a waste of time. The Wednesday meeting in London brought together the Communication Workers Union (CWU), Unite the Union, a third-party arbitrator, and DeepMind HR representatives. Missing from the room: any senior DeepMind leadership.

That absence set the tone. "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," said John Chadfield, a CWU officer who attended. "Negotiations have stalled at an early stage."

DeepMind disputes the characterization. "The first step in the process is to define who the unions want to represent and the parties agreed on next steps to do this," said Al Verney, a Google DeepMind spokesperson. "The appropriate representatives attended this initial meeting."

What triggered the unionization push?

The organizing effort traces back to February 2025, when Alphabet quietly removed a pledge from its ethics guidelines stating it would not use AI for weapons development or surveillance. For researchers who joined DeepMind specifically because of those principles, the change felt like a betrayal.

"Those principles were a big part of why I joined DeepMind," one employee told WIRED anonymously. "We basically just got rid of them all."

The ethics pivot came alongside a broader shift. In April, The New York Times reported that Google had struck a deal allowing the Pentagon to use its AI for "any lawful government purpose." Roughly 600 US-based Google employees signed a letter protesting the terms. The Department of Defense later confirmed deals with seven AI companies, including Google, OpenAI, Microsoft, and SpaceX, to deploy their models on classified networks.

What happened in the meeting itself?

During Wednesday's session, a DeepMind employee read a prepared letter on behalf of colleagues supporting unionization. According to multiple sources, HR representatives interrupted the reading twice.

The letter alleged that Google has tried to suppress internal discussion about unionization. It claims the company shut down or reconfigured internal chat venues and blocked staff from responding to company-wide communications about the union bid. Employees who tried workarounds were "reprimanded" by HR, the letter states.

“Instead of having meaningful dialog with its employees about our concerns, Google DeepMind workers have been treated as a problem handed off to HR.”

— Letter read by DeepMind employee at union meeting

"The intention was to intimidate," said one employee involved in drafting the letter. "These are well-established union-busting techniques."

Google's response was measured. "We'll continue to engage constructively in the process and have open dialog with employees," Verney said. "For topics outside of this, we continue to offer employees a variety of other channels and opportunities to discuss their views."

Why UK labor law matters here

This fight is playing out in London, not Mountain View, and that distinction is significant. UK labor law provides unions with legal pathways to recognition that simply do not exist in the United States. The Communication Workers Union and Unite are established organizations with decades of collective bargaining experience.

In May, DeepMind employees formally asked Google to recognize the two unions as joint representatives. Google denied the request but agreed to participate in negotiations through a third-party arbitrator. If those negotiations collapse, the unions can petition an arbitration committee to compel recognition.

"We're hoping that Google genuinely comes to the table and we can agree something amicably," Chadfield said.

The broader context for AI worker organizing

DeepMind is not alone. In late February, staff at DeepMind and OpenAI signed an open letter supporting Anthropic after the Department of Defense sought to designate the lab a supply chain risk. Anthropic's offense? Refusing to allow its technology in autonomous weapons or mass surveillance systems.

The Alphabet Workers Union, formed by US-based Google employees in 2021, has roughly 1,400 members. Alphabet does not recognize it for collective bargaining, but the union has successfully negotiated agreements on behalf of Google contractors.

For DeepMind's London staff, the stakes feel personal. Many joined the lab because of its stated commitment to beneficial AI. Watching those commitments erode while being told to take concerns to HR, not leadership, has pushed them toward formal organizing.

Frequently Asked Questions

Why are DeepMind employees trying to unionize?

The effort began after Google removed ethics pledges about not using AI for weapons and surveillance. Employees also cite concerns about military contracts and internal suppression of dissent.

Has Google ever recognized a union?

No. Alphabet has never formally recognized a union for collective bargaining in the US or UK. The Alphabet Workers Union exists but operates without formal recognition.

What unions are involved in the DeepMind push?

The Communication Workers Union (CWU) and Unite the Union are seeking joint recognition to represent London-based DeepMind employees.

Can UK unions force Google to recognize them?

Yes. Under UK law, if negotiations fail, unions can petition the Central Arbitration Committee to compel recognition if they can demonstrate sufficient worker support.

Source: Feed: Artificial Intelligence Latest / Joel Khalili