About 300 UK staff at Google-owned DeepMind are seeking to unionise in order to challenge Google’s decision to sell AI technology to military organisations and its ties to the Israeli government, the Financial Times reported.

The staff have reportedly sought to join the Communication Workers Union in recent weeks after Google in February dropped a promise not to develop AI for weaponry or surveillance purposes.

The employees were also concerned about media reports of Google selling cloud services and AI technology to Israel’s Ministry of Defence, the report said.

Reversed commitment

The Israeli government has a $1.2 billion (£890m) cloud computing agreement with Google and Amazon called Project Nimbus.

Media reports said the Israel Defence Forces have used AI systems to generate targets for assassinations and attacks in the Gaza strip, although it was unclear whether the IDF was using commercially purchased software or custom-built tools for the attacks.

An engineer involved in the unionisation effort said the employees were “putting two and two together” and believed the technology they were developing was being used in the Gaza conflict.

“This is basically cutting-edge AI that we’re providing to an ongoing conflict. People don’t want their work used like this,” the engineer said, adding, “People feel duped.”

Five DeepMind staff have quit over the past two months citing the Israel cloud deal and Google’s reversal of its commitments around the harmful use of AI.

In the US Google fired some staff last year for sit-in protests over Project Nimbus.

In May 2024 DeepMind staff sent a letter to company leadership calling it to drop military contracts and held meetings with leaders, but their requests have reportedly been denied.

Staff discontent

The unionisation effort requires a vote by DeepMind’s roughly 2,000 employees in the UK to be formally recognised by the company.

Google said it has always developed and deployed AI “responsibly” and that it encourages constructive and open dialogue with employees.

It said the landscape has changed significantly since its 2018 promise not to sell AI for weapons or surveillance.

In 2018 several Google staff quit and thousands of employees signed a petition in protest against Project Maven, a US military contract using AI technology to improve drone strikes.

Following the discontent Google did not renew the Pentagon contract and made its pledge not to develop AI technologies that “cause or are likely to cause overall harm”, including weapons and surveillance.