Workers at Google DeepMind. Google’s AI research laboratory. Have voted to unionize, in part due to concerns about a recent deal with the US defense department; In a letter to management obtained exclusively by The Guardian, staff requested recognition of the Communication Workers Union and Unite the Union as their representatives.

Concerns Over AI and Military Applications

One of the workers, who spoke on condition of anonymity, cited reports that Google was close to finalizing the deal as a key motivation for unionizing. The worker referenced the US military’s actions in Iran and the Trump administration’s conflict with Anthropic as signs that the department might not be a responsible partner. The agreement was officially announced on Friday.

“I have joined the union due to concerns about AI being used to empower authoritarianism, whether through military or surveillance applications, both foreign and domestic,” the worker said. “By unionizing, we are taking the traditional route for workers to organize and have a say.”

Complicity in Israel’s War in Gaza

Another DeepMind worker, also anonymous, said many colleagues were troubled by what they saw as their complicity in Israel’s war in Gaza. The company has provided the Israeli military with increased access to its AI tools, according to a 2023 report by The Washington Post. In 2021, Google signed a $1.2 billion cloud-computing contract with the Israeli government, alongside Amazon.

“Our technology helped the IDF,” the worker said, referring to Israel’s military. “I want AI to benefit humanity, not to facilitate a genocide.”

Google did not respond to a request for comment.

Concerns among Google workers and investors have been building for years but intensified after the company dropped a pledge not to develop militarized AI last year. The decision was a key factor in the UK workers’ decision to seek unionization, two of them said.

While some Google employees in the US have unionized in the past, the UK DeepMind workers are the first in a “frontier” AI lab to pursue union recognition, they noted. Google DeepMind is based in London but has offices across North America and Europe. At least 1,000 workers would be represented if the company recognizes the union, according to union officials.

Broader Industry Trends and Investor Pressure

The Pentagon confirmed on Friday that it had reached agreements with seven leading AI companies, including Google, SpaceX, OpenAI, Nvidia, Microsoft, Amazon Web Services, and Reflection. Anthropic, whose technology is widely used by the US military, was notably absent from the list.

“These agreements accelerate the transformation toward establishing the United States military as an AI-first fighting force and will strengthen our warfighters’ ability to maintain decision superiority across all domains of warfare,” defense department officials said in a statement.

The Trump administration has pushed AI companies to make their tools available on classified networks without the usual restrictions. Google’s contract reportedly includes a clause stating that the AI system is not intended for domestic mass surveillance or autonomous weapons without human oversight. However, the clause is non-binding, and the agreement also states that Google has no right to control or veto “lawful” government decision-making.

Workers who voted to join the union said they aim to pressure Google to meet demands already raised by other employees. These include a commitment not to develop technology “whose primary purpose is to cause harm or injury to people,” the establishment of an independent ethics oversight body, and the right for workers to refuse to contribute to projects on moral grounds. If Google refuses, the workers are considering protests and “research strikes,” in which staff would stop working on major projects like Gemini, Google’s AI bot, while continuing less impactful updates.

More than 600 Google employees signed an open letter last week to CEO Sundar Pichai, urging the company not to make its AI systems available for classified use. “We want to see AI benefit humanity; not to see it being used in inhumane or extremely harmful ways,” the letter stated. “Making the wrong call right now would cause irreparable damage to Google’s reputation, business, and role in the world.”

Tech workers have increasingly challenged their companies over the use of AI in military contexts. In 2024, Google terminated 50 employees who had protested against Project Nimbus, a 2021 contract with the Israeli government. Microsoft, which was revealed to have supplied Israel with cloud storage for mass surveillance of Palestinians, faced worker protests and eventually ended the military’s access to that technology.

Investors have also raised concerns. A group of shareholders owning about $2.2 billion in Alphabet shares wrote a letter to Google’s parent company, demanding greater transparency about AI deployments in “high-risk” contexts. The letter cited concerns over the company’s work with US immigration authorities and Project Nimbus, as well as questions about the effectiveness of internal policies and Board oversight in conflict-affected or security-sensitive environments.

Google has faced similar employee protests in the past. In 2018, employees widely protested a military contract known as Project Maven, in which the company agreed to build AI tools for the Pentagon’s analysis of drone footage. In response to the backlash, Google did not renew the contract in 2019 and published a set of AI principles that included a pledge not to design AI for weapons. Palantir later took over the project, which continues today.