Hundreds of employees at Google have called on chief executive Sundar Pichai to block the use of the company’s artificial intelligence tools in classified military operations, highlighting growing tensions within the tech industry over the role of AI in defence.
In an open letter signed by more than 560 staff, including engineers and researchers across AI and cloud divisions, employees urged the company to refuse contracts that could allow its technology to be used for purposes such as autonomous weapons or mass surveillance. They argued that participation in classified projects could limit oversight and increase the risk of misuse.
“We want to see AI benefit humanity, not being used in inhumane or extremely harmful ways,” read the letter, which was sent to Pichai yesterday. “This includes lethal autonomous weapons and mass surveillance, but extends beyond.”
The letter further mentions, “The only way to guarantee that Google does not become associated with such harms is to reject any classified workloads. Otherwise, such uses may occur without our knowledge or the power to stop them.”
The campaign comes at a time when Google is reportedly working with the US Department of Defense to use its artificial intelligence (AI) models, such as Gemini, for classified activities.
Another anonymous participant in the campaign said, “This isn’t just about the military. AI-powered mass surveillance is a direct threat to American civil liberties.” “This is not low-risk and theoretical; we are already in these fights. We see AI being used to support authoritarianism in China.”
The matter becomes even more significant as the letter gets signed by more than 18 senior Google employees including VPs and directors, highlighting the depth of the matter.
The employee letter reflects a wider discussion in the tech industry about the ethical implications of artificial intelligence. The letter warned that the use of AI systems in classified environments could conceal their use and lead to a loss of accountability. These concerns are not unique, with others raising issues about the use of AI in lethal autonomous weapons and mass surveillance programs.
The issue has gained urgency following a dispute between the Pentagon and Anthropic, which refused to grant unrestricted access to its AI models for military use. The standoff led to the company being labelled a supply-chain risk by US authorities and removed from certain government contracts.
Employees Push for Clear Ethical Boundaries
Google and other AI companies’ employees have increasingly spoken out about the use of their technologies.
In January, hundreds of employees from Google and OpenAI called for restrictions on the use of AI in military contexts, including the development of domestic surveillance systems and autonomous weapons systems.
This letter builds on that momentum, arguing that the only way to prevent AI systems being used for harmful purposes is to refuse to work on classified military contracts.
Growing Pressure on Tech Companies
The move also echoes Google’s 2018 decision to pull out of Project Maven, a US Department of Defense project that applied AI to analyse images from drones, after employee protests.
Since that time, the company has upheld AI principles focused on responsible development, but recent events indicate these principles are being reassessed as governments look to leverage advanced AI for national security.
The debate highlights the increasing pressure on major technology firms to define their stance on military AI.
States are moving faster to incorporate advanced AI systems, and companies are caught between the need to collaborate with governments to support national security and the need to maintain ethical principles.
For now, the employee response signals that internal resistance remains strong, even as commercial and strategic incentives for collaboration with defence agencies continue to grow.
Why this matters
This is more than just Google and one memo. It’s part of a broader trend in the global positioning of AI.
Over the last few years, AI has been primarily viewed as a business tool – to boost productivity, automate processes and create products for consumers. This is now shifting to a more nuanced story. AI is now seen as critical infrastructure. More than just code, but something that can affect defence, intelligence and security.
That makes things more complicated for businesses. One is the pressure to work with governments, particularly in areas of national interest. On the other, there is increasing pushback from employees who are concerned about the use of these technologies and whether there are sufficient protections in place.
The Google letter is at the crossroads. It highlights a pragmatic issue. When AI systems are used in classified settings, transparency decreases. Visibility into how they are used, changed or extended is reduced. This introduces a risk that companies may not be able to manage after deployment.
This is highlighted by the Anthropic case. In not allowing unfettered access and wanting to impose guardrails, the company set a line in the sand. Its reaction – being declared a supply-chain risk – demonstrates the difficulty of doing so when national security is on the line.
Here’s the rub. AI is powerful enough to be useful in defence, but it’s also difficult to track. This is a message to investors and watchers – AI is no longer just a technology cycle, it’s becoming a geopolitical weapon.
And as this occurs, choices about access, control and use will impact not just firms, but markets and regulation too.
