Google DeepMind engineer Sofia Liguori resigns over military AI contracts

The former research engineer has criticized Google's military and surveillance partnerships, alleged a culture of workplace intimidation and called for stronger employee action over the use of AI in warfare

Google DeepMind engineer Sofia Liguori has resigned following objections to Google's military and surveillance contracts, raising concerns about AI ethics, corporate oversight and employee accountability

Google DeepMind research engineer Sofia Liguori has resigned from the company, citing objections to Google's expanding military AI partnerships and what she describes as a workplace culture that discourages employees from challenging how its technology is deployed.

Liguori confirmed her departure in an open letter dated October 20, 2026, following more than three years at DeepMind. She had been on leave since Google signed a classified agreement with the US Department of War in April, having previously participated in employee activism opposing the company's military and surveillance contracts.

Her resignation centers on concerns about the integration of AI into military operations, the extent of Google's oversight over how its technology is used and the responsibilities of employees developing systems that may have applications beyond their original purpose.

In her letter, published in a personal capacity, Liguori wrote:

"I no longer believe that our technology will have a net positive impact on society."

Military AI contracts prompt resignation

Liguori identifies Google's classified Department of War agreement as the decisive factor behind her departure, drawing connections between the contract and Project Nimbus, the company's cloud computing arrangement involving the Israeli government.

She argues that the structure of these agreements allows AI technology to be integrated into military and surveillance systems without sufficient restrictions on its eventual use.

In a LinkedIn post sharing her resignation, she wrote:

"The modus operandi tested in Gaza with Nimbus was legitimised as the standard with April's DoW classified deal: contracts that allow the integration of AI into military and surveillance systems while imposing no restrictions on permissible uses and leaving the company no monitoring power."

These are Liguori's characterizations of the agreements and their implications. Her letter also raises objections to Google's relationships with Lockheed Martin and US agencies including the Department of Homeland Security, Immigration and Customs Enforcement and Customs and Border Protection.

A central concern is whether AI companies can uphold meaningful ethical commitments when they have limited visibility or control over how customers ultimately deploy their technology.

In response to these claims, a Google spokesperson told the BBC: "We work with public sector organizations worldwide and follow a consistent process for reviewing and entering these contracts, including compliance with our acceptable use policies and AI principles.

"We also have dedicated channels for employees to safely raise any concerns - ranging from business ethics to misconduct - and have strict workplace policies in place to prevent retaliation of any kind."

Liguori argues that discussions about AI safety have become too narrowly focused on models behaving unpredictably or acting outside human control, rather than examining the consequences of their intended applications.

She questions whether making AI systems more reliable necessarily makes their deployment acceptable, particularly when the systems are developed for military targeting or surveillance.

Engineer challenges claims that technology is neutral

Liguori joined Google DeepMind in May 2023, working as a Research Engineer in optimal power flow, energy and sustainability.

Her objections extend beyond the military contracts to the wider development of general-purpose and agentic AI, particularly the concentration of technological and political influence among a small number of companies.

In her departure letter, she wrote:

"Technology is NOT a neutral tool. This assessment assumes we can separate technology from the context it's deployed into, but we should not: we have a responsibility to assess the tool in the context in which it's going to be used."

She distinguishes between task-specific AI applications, including the grid optimization research she worked on, and the increasingly capable general-purpose models being developed and deployed across society.

Her criticism also extends to the possible consequences for education, employment and how people learn and interact, alongside concerns about surveillance, misinformation and democratic accountability.

For Liguori, responsibility does not end with the technical development of a system. She argues that researchers and engineers must also consider who controls the technology, who benefits from its deployment and the purposes for which it is being used.

Allegations of workplace intimidation and calls for employee action

Beyond Google's commercial decisions, Liguori alleges that employees face internal pressure not to discuss controversial partnerships openly.

Her letter describes a culture in which military contracts are treated as subjects employees should avoid raising, including within teams whose work may have implications for those agreements.

She claims that challenging these boundaries resulted in meetings with HR over alleged internal policy breaches, which she characterized as intimidation rather than substantive disciplinary action.

"We also need to strongly reject this corporate culture of censorship and intimidation," she wrote.

Liguori also criticizes what she regards as a tendency among technology workers to separate their individual projects from the wider consequences of their employer's activities.

Her resignation follows a period of activism that included taking time away from work and demonstrating outside the office. Although she has decided to leave DeepMind, she argues that collective employee action is more effective than individual departures in influencing how AI companies operate.

Her letter directs current employees toward several organizing efforts, including Google DeepMind workers involved with UTAW and Unite, No Tech for Apartheid, Googlers Against ICE and the cross-company Coalition of Concerned AI Staff.

The demands she identifies include enforceable restrictions on military and surveillance applications, greater contract transparency, the right for employees to abstain from work on ethical grounds and stronger sustainability commitments.

Liguori closes by encouraging employees who remain at Google DeepMind to organize collectively, arguing that technology workers retain considerable influence over how AI is developed and deployed even when they disagree with the direction taken by company leadership.

ETIH Innovation Awards 2027 open for entries

The ETIH Innovation Awards 2027 are now open for entries, with 26 categories recognising companies, institutions, projects and individuals making a measurable impact across education technology. The global awards span AI in education, digital learning, higher education, skills and workforce, inclusion, education operations, evidence and impact, with entries managed through ETIH’s new online awards platform.

Entries are open worldwide, with the early-bird deadline on Friday 18 December 2026 and the final entry deadline on Friday 26 February 2027 at 23:59 UK time. 

Explore the 2027 categories and judging criteria, read the How to Enter guide or start an entry through the ETIH Innovation Awards 2027 platform.