Unions seek UK ruling on DeepMind union ballot after recognition talks stall

Google DeepMind’s UK union drive has moved beyond talks with management. On November 12, the Communication Workers Union and Unite applied to a government arbitration body to establish support for a unionization ballot, WIRED reported. The workers’ ambitions extend beyond workplace terms: they want a collective voice in how the lab’s AI is developed and deployed.

From voluntary talks to a binding ballot

The application went to the Central Arbitration Committee, which resolves disputes between employers and unions. WIRED reviewed the filing. It asks the committee to confirm that the two unions, chosen by employees as joint representatives, have enough worker support to trigger a vote. Google DeepMind would be legally required to respect the ballot’s result.

If the effort succeeds, roughly 1,000 staff in product-focused roles would qualify for union membership. Organizers expect a ballot in the coming months, according to an email sent to all UK staff on Thursday and seen by WIRED. That timing is the employees’ expectation, not an announced committee schedule.

The filing follows meetings among employees, union representatives and the company about voluntary recognition. John Chadfield, the CWU’s national officer for technology, accused management of treating the process as an administrative exercise handled by HR and labor-relations staff. Google DeepMind did not immediately respond to WIRED’s request for comment.

We want a union, we want collective bargaining, and we want a say in how AI is developed and deployed

DeepMind employees organizing the union drive, in an email to UK staff seen by WIRED

Military work is a central grievance

The organizing push began in February 2025, when Google’s parent company, Alphabet, removed an ethics pledge against using AI for purposes such as weapons development. That change remains central to the employees’ case for organizing: they describe the effort as a way to restore ethical boundaries they believed were part of joining DeepMind.

One employee told WIRED that DeepMind had emphasized its principles and “red lines” when they joined. The employee said much of the union effort was about reestablishing that ethical framework. Both employees interviewed by WIRED requested anonymity because they feared backlash, underscoring the sensitivity of challenging the company’s direction from inside the lab.

Organizers specifically object to Google’s agreements with the US and Israeli militaries. If they unionize, they intend to demand that Google withdraw from those contracts. Their aim therefore reaches beyond securing representation: they want to use collective bargaining to press the company over who receives its technology and what that technology supports.

Google has previously defended its national-security work. In an April statement to The New York Times, quoted by WIRED, spokeswoman Jenn Crider said Google was proud to provide AI services and infrastructure supporting national security. She also said the company opposed domestic mass surveillance and autonomous weaponry without appropriate human oversight.

An organizing model for other labs

The employees also hope their campaign can provide a playbook for UK staff at other major AI labs, including OpenAI and Anthropic. One organizer argued that an industry with such economic power gives employees and regulators too few opportunities to provide checks and balances. Collective organization, the employee said, could let staff articulate a shared vision for how their technology should be used.