DeepMind Researchers Quit Google in Protest of Military A.I. Deals
Some Google employees believe the company has moved away from its old motto, “Don’t be evil.” Several workers at Google DeepMind, Google's artificial intelligence (AI) division, have quit or are getting ready to leave because they oppose Google's work with the U.S. military and Israel, the BBC reports. These employees doubt that Google can control how its AI models are used in war.
‘A Net Positive Impact’ No Longer
One employee who is leaving is Sofia Liguori, a research engineer at Google DeepMind in London. She is Italian and joined the company in 2023. Her work focused on using AI to send electricity through power grids more efficiently. In an internal farewell letter quoted by the U.S. news outlet The Intercept, she wrote, “I no longer believe that our technology will have a net positive impact on society.” In other words, she no longer believes Google's technology will do more good than harm for society.
She objects to two projects. One is Project Nimbus, a deal to provide cloud computing and AI services to the Israeli government and military. The other is a Pentagon contract that allows Google's AI models to be used for secret military work. The day after the Pentagon deal became public in April, she called in sick and took vacation. Ms. Liguori said her efforts to protest inside the company did not lead to change. Posts that included the word “genocide” were deleted. Managers and human resources staff said her tone was too aggressive, and complaints received replies that were only partly automated.
The BBC reports that AI researcher Andreas Kirsch also left DeepMind earlier this month. In a series of posts on X during the summer, he argued that all the conditions set when Google bought DeepMind in 2014, including a rule against military use, had disappeared. “Not one safeguard survived intact,” he wrote. “I learned about it from the press.”
The Pentagon Deal
In April, Google reached an agreement with the U.S. Department of Defense. It allows the military to use Google's AI models for “any lawful government purpose.” The contract says the models should not be used for mass surveillance of people inside the United States or for weapons that operate on their own without proper human oversight. However, Google cannot block military decisions. Critics therefore believe these limits may be little more than promises about how the technology should be used.
Google is not the only company making these deals. OpenAI and xAI have also signed contracts with the Pentagon. Anthropic, however, was placed on a Defense Department blacklist after it refused to remove safeguards meant to prevent its AI models from being used for weapons and surveillance.
Google Defends Itself
Google disagrees with the criticism. Company spokesman Tom Hewett said Google works with public sector organizations around the world and reviews all contracts using the same process, which includes its AI principles. He added that employees can raise concerns inside the company and that punishing employees for doing so is not allowed. In April, Google had already said it believed that giving the government access to its commercial AI models under standard industry terms “represents a responsible approach to supporting national security.”
Not a New Conflict
The disagreement has been going on at Google for years. In 2018, thousands of employees protested Project Maven, a Pentagon project involving drones. About a dozen employees quit, and Google eventually left the project. In 2024, the company fired about 50 workers after they protested Project Nimbus inside its offices.
This time, employees are organizing their protests more closely. According to The Intercept, DeepMind employees voted to form a union in May. In February, workers also protested outside the company's London office, using the slogan “No AI in the kill chain.” Ms. Liguori believes these actions are not enough. She thinks employees should have gone on strike this year.