Google is facing strong opposition from some of its employees over the application of its artificial intelligence technology for military purposes.
The BBC spoke with several tech workers—some of whom have recently resigned from Google DeepMind, the company’s AI division, while others are currently in the process of leaving—due to concerns regarding the tech giant’s military contracts.
Among them is Sofia Liguori, a London-based research engineer. She resigned from Google DeepMind this week and shared her concerns exclusively with the BBC, despite contractual restrictions that would otherwise prevent her from doing so.
“Realizing that my company is increasingly pushing AI systems for use in military operations makes me ashamed to work at Google,” Liguori said, speaking in a personal capacity.
In response, a Google spokesperson told the BBC: “We work with public sector organizations around the world and follow a consistent process to review and enter into these contracts, which includes adherence to our Acceptable Use Policy and AI Principles.”
“We also have dedicated channels for employees to safely raise any concerns—ranging from business ethics to misconduct—and have established strict workplace policies to prevent any form of retaliation.”
Earlier this year, Google was among several technology companies that signed contracts with the U.S. military to deploy artificial intelligence within classified military environments. U.S. Secretary of Defense Pete Hegseth has identified the integration of AI into the military as a key priority, stating that the Trump administration will build an “AI-first” fighting force.
However, as artificial intelligence technology develops rapidly, critics question whether AI is sufficiently reliable for use in warfare—especially without more extensive testing, robust safeguards, and human oversight.
Liguri, who spent three years at Google, stated that she decided to resign due to the company’s growing collaboration with the U.S. and Israeli militaries. “I do not align with this direction, and it has caused me to lose faith in the company’s ability to uphold the ethical values it claims to follow,” she said.
According to Google’s own AI principles, these values include ensuring “human oversight” in the implementation of AI and aligning the technology with “international law and human rights.” Yet, both former and current Google employees—as well as experts interviewed by the BBC—have expressed concerns regarding the adequacy of human oversight at critical junctures.
Tech workers have expressed particular concern regarding artificial intelligence military agreements signed this year, as these deals significantly expand the scope of the Pentagon’s use of such technology.
AI researcher Andreas Kirsch also resigned from Google DeepMind earlier this month, citing his concerns.
Kirsch told the BBC that he believes “Google’s governance and oversight are insufficient to safely and responsibly deploy AI in military and classified settings,” or to “safely develop AI in the future.”
He added: “Google does not appear to have the authority to control or veto how the Pentagon uses its AI, nor does it seem to have sufficient insight into how that AI is being deployed.”
Over the past few years, the U.S. military has deployed a variety of artificial intelligence technologies to help enhance its operational capabilities.
The most prominent example is the Maven system—a primary software platform used by the Pentagon to organize intelligence and data.
Developed by the U.S. defense firm Palantir, the Maven system leverages artificial intelligence to assist personnel in selecting and identifying targets for strikes. Palantir has integrated technology from several leading AI companies into the Maven system, significantly improving the efficiency of certain processes.
Retired Lieutenant General Jack Shanahan, who formerly headed “Project Maven,” noted that the project’s “philosophy” appeared to be: “deploy [AI technology] as broadly and as quickly as possible—potentially at the expense of rigor and discipline in testing and evaluation.”
Nevertheless, the Pentagon has consistently maintained that adopting artificial intelligence is essential for winning wars. Earlier this year, Emil Michael, a U.S. Department of Defense official, stated: “In the age of AI, speed determines the winner.”
Earlier this year, Google signed a contract with the Pentagon to allow its artificial intelligence technology to be used for classified military work. The agreement stipulates that Google’s AI models may be used for “any lawful government purpose,” but specifies that they “may not be used for large-scale domestic surveillance or for autonomous weapons—including target selection—without appropriate human oversight and control.”
OpenAI, Amazon, Microsoft, SpaceX, Oracle, Nvidia, and the startup Reflection have also signed similar contracts, enabling the deployment of artificial intelligence technology across various military contexts—such as logistics, planning, and battlefield operations.
As the technology advances rapidly, there is growing concern regarding the use of AI in warfare.
“AI compresses the time available for operators to make decisions, as it operates far faster than a human—a factor that is particularly dangerous in high-pressure environments,” says Sarah Shoker, former lead of the geopolitics team at OpenAI.
“When an AI selects hundreds of targets in a very short span of time, it becomes virtually impossible for human operators to meaningfully review the AI’s output or identify errors.”
Google employees told the BBC that they feel unable to speak openly about AI safety and its application in warfare, as their employment contracts prohibit them from publicly criticizing the company.
Sophia Liguori told the BBC, “In principle, I am barred from criticizing Google for the rest of my life.” Yet, she chose to speak out, explaining, “I believe it is important for people to be aware of these issues, and I don’t consider stating facts to be the same as criticism.”
Liguori added that at Google, “people fear retaliation and the potential damage to their careers. You feel as though you are being watched, which makes you afraid to speak up.”
An employee currently working at Google DeepMind—who requested anonymity for fear of retaliation—told the BBC, “Many of the highly motivated employees who joined the company early on may end up leaving—but there is a fresh wave of new recruits ready to fill those gaps. It could represent a shift in power.”
Another Google DeepMind employee, who wished to remain anonymous, told the BBC: “Many people have left the company because they were deeply unhappy with Google DeepMind’s shift in its ethical stance.”
“Some chose to push back, while others took the view: ‘It’s a big corporation—what did you expect to change?'” the employee said.
Google’s agreements involving Israel have also sparked controversy. In 2024, Google faced criticism from its own staff over its participation in “Project Nimbus”—a $1.2 billion (£960 million) contract with Amazon to provide artificial intelligence technology and cloud computing infrastructure to the Israeli government and military. At the time, Google stated that “Project Nimbus is not directed at highly sensitive, classified, or military workloads relevant to weapons or intelligence services.”
Furthermore, the use of AI capabilities by Israel during the war has become a point of contention. A recent documentary titled Naza explores the role of AI-powered targeting systems in the bombing of the Gaza Strip, which has resulted in a high number of civilian casualties. An Israeli government minister has dismissed the claims made in the film as untrue and a distortion of the facts.
Andreas Kirsch, a former Google employee, argues that AI technology is not yet sufficiently mature for use on the battlefield. “Current AI models are not powerful enough, and I don’t believe they have been trained for these kinds of tasks,” he said.