Former Google DeepMind engineer Sofia Liguori has accused the tech giant of helping legitimize the Israeli regime’s criminal actions against Palestinians through its $1.2 billion Project Nimbus contract, revealing how the company’s artificial intelligence tools are made available to Israeli government bodies amid the genocide in Gaza.

Liguori, who resigned from Google after repeatedly challenging its cooperation with Israel, said she could no longer reconcile her work on supposedly humanitarian AI projects with the company’s involvement in providing technology to the Israeli regime.

“I felt so hypocritical that we were just talking about all these good things and not even mentioning Project Nimbus,” she told openDemocracy, describing her former work as “building the nice face of the devil”.

Project Nimbus, signed by Google and Amazon Web Services with Israel in 2021, provides cloud infrastructure and data storage services to Israeli public institutions. The agreement reportedly prevents the companies from shutting down their services in Israel or denying access to particular government entities.

Despite Google's claims that its services are not intended for highly sensitive military operations, leaked training materials revealed that Israeli government departments, including the Ministry of Defence, have access to Google Cloud's artificial intelligence capabilities, including facial recognition, object tracking and sentiment analysis.

According to investigations by The New York Times and the Centre for Constitutional Rights, Israeli authorities have used facial recognition and surveillance technologies to identify and monitor Palestinians in the occupied West Bank, East Jerusalem and Gaza.

Liguori said the contradiction became impossible to ignore during her maternity leave, when she watched the unfolding genocide in Gaza while knowing that her employer was involved in the technology infrastructure supporting the Israeli regime.

“I watched a live-streamed genocide for a year, knowing that my company was complicit,” she said.

After returning to work in August 2025, she urged colleagues to confront the company's involvement in Project Nimbus, arguing that opposing genocide should be central to any genuine commitment to developing AI for the public good.

However, management rejected her request to discuss the project in team meetings. She also said internal links to reports containing the word “genocide” were removed for allegedly violating company policies against aggressive language.

Google's expanding military partnerships

The dispute intensified as Google expanded its military-related business.

In 2018, following an employee revolt involving approximately 3,000 workers, Google announced it would not renew its Pentagon contract for Project Maven, which involved artificial intelligence technology for military drone operations.

The company subsequently introduced AI principles prohibiting the use of its technology for weapons and human rights abuses.

However, Google later removed references to weapons and surveillance from its AI principles, while expanding partnerships with military contractors, including Lockheed Martin and Palantir.

In April 2026, Google also signed an agreement allowing the US Department of War to use its AI technology for government purposes, including mission planning and weapons targeting.

Liguori said a petition signed by more than 600 employees opposing the Pentagon agreement failed to change the company's direction.

“I reached a level of disgust with my company that was just incompatible with me continuing to stay there,” she said.

She officially resigned on October 20, publishing a letter in which she questioned whether the development of general-purpose AI could deliver meaningful benefits to humanity under the prevailing corporate and military priorities.

Collective action forces Google to retreat

Liguori's departure also exposed Google's attempts to silence former employees through lifetime non-disparagement clauses in employment agreements.

Despite the legal risks, she publicly discussed the company's dealings with Israel and published her resignation letter, arguing that the public had a right to know how major technology companies were contributing to military and surveillance systems.

Following pressure from employees, trade unions and the No Tech for Apartheid campaign, Google informed its UK workforce this week that it would not enforce post-employment non-disparagement provisions against former employees criticizing the company or its AI products.

Liguori attributed the decision to collective action by workers, the United Tech and Allied Workers union and campaign organizations.

Google said it had updated its terms for new employees, removing the post-termination element of non-disparagement clauses.

For Liguori, the decision represents a victory against corporate attempts to suppress dissent and prevent employees from exposing the ethical consequences of their employers' activities.

“It's also a sign that our collective effort had an effect, and that it's worth challenging the rules,” she said.

Google had not responded to openDemocracy's request for comment at the time of publication.