DeepMind staff vote to unionise over Pentagon AI deal and military ties
Employees cite concerns regarding the US Department of Defence as a responsible partner and demand the right to refuse projects on moral grounds.
DeepMind workers at Google's London headquarters have voted to unionise, formally requesting recognition from the Communication Workers Union and Unite the Union. The ballot, held in April, marks a significant shift for the team behind Google Gemini and other unified AI technologies, driven largely by anxieties surrounding a new commercial agreement between the company and the US Department of Defence.
The primary catalyst for the industrial action appears to be the Pentagon's recent deals with leading AI firms, which grant the defence department the right to utilise these technologies for any lawful purpose. Employees have expressed strong reservations about the US government acting as a responsible partner, pointing specifically to the department's involvement in the Iran conflict and its ongoing disputes with rival artificial intelligence firm Anthropic as evidence of potential instability.
Beyond concerns regarding US military engagement, workers have voiced specific alarm over the potential deployment of their technology by the Israel Defence Forces. This fear is compounded by the knowledge of Google's existing $1.2 billion cloud computing contract with the Israeli government, signed in 2021. The union aims to leverage this situation to pressure Google into committing against developing technology whose primary purpose is to cause harm or injury to people.
In response to these concerns, the union is seeking to establish an independent ethics oversight body to monitor the development and application of these powerful models. Furthermore, employees are demanding the explicit right to refuse to contribute to specific projects on moral grounds, arguing that current contractual frameworks do not provide sufficient protection for staff involved in high-risk defence applications.
The move comes in the wake of previous tensions within the tech sector, including Google terminating more than 50 staff members in 2024 following a protest regarding the company's military ties to the Israeli government. Additionally, in late February, staff at DeepMind and OpenAI signed an open letter supporting Anthropic following scrutiny of Pentagon deals, a move that preceded the removal of an ethical pledge by the companies involved.
Management has been given ten working days to voluntarily recognise the union before legal processes are initiated. The outcome of this bid will determine whether DeepMind staff can enforce their moral objections to military contracts and ensure that the artificial intelligence they build is not deployed without appropriate human oversight and control.
