Google Whistleblower Resigns Over Reported Gemini Deployment for Classified Military Work; Press Release Confirms Only Unclassified Phase

A Google DeepMind research engineer has resigned in protest and gone public with concerns that the company is expanding its military AI contracts beyond what it has previously acknowledged. Sophia Liguori, 32, says she left the company on October 20, 2026 after failing to convince executives to end Google's involvement in Israeli military operations and new Pentagon partnerships, according to reporting by openDemocracy and techpolicy.press [1][2].

Liguori worked in Google DeepMind's London-based Impact Accelerator team, which focused on sustainability applications such as optimizing electricity grid power flow. She says she was unaware of Project Nimbus—a $1.2 billion Google and Amazon Web Services contract with Israel signed in 2021—when she joined in May 2023 [1][2].

Project Nimbus and the policy shift

Project Nimbus provides cloud infrastructure to Israeli government entities, with contractual terms preventing Google or AWS from denying service to any government body. Google has stated the work is "not directed at highly sensitive or classified military workloads relevant to weapons or intelligence services" [1][2]. However, training materials leaked to The Intercept in 2022 showed the full suite of Google Cloud AI tools—including face detection, object tracking, and sentiment analysis—were available to Israel's Ministry of Defense [1][2].

In early 2025, weeks after Donald Trump's second term began, Google updated its AI Principles to remove prohibitions on weapons and surveillance applications. DeepMind chief Demis Hassabis and research labs senior vice president James Manyika wrote that "democracies should lead in AI development, guided by core values like freedom, equality, and respect for human rights" [1][2].

Lockheed Martin announcement: reported classified use vs. confirmed unclassified phase

openDemocracy reports that Google announced its Gemini models would be used for "both classified and unclassified purposes at the AI factory of Lockheed Martin," a primary U.S. defense contractor building military aircraft, missiles, and satellites [1][2][4]. Liguori's own account addresses other expansions, such as a June 2026 partnership with military contractor Palantir [1][2].

A Google Cloud press release from October 29, 2025 confirms a "strategic collaboration" to integrate Gemini into the Lockheed Martin AI Factory via Google Distributed Cloud, deployed in "secure, on-premises, and air-gapped environments" [5]. The release specifies that the first phase covers an "unclassified on-premise environment." It does not mention classified deployment [5].

Liguori told openDemocracy: "You never know then how your technology will be used in Palantir's systems," referencing the June 2026 expansion of Google's partnership with Palantir [1][2].

Pentagon contract and internal dissent

According to openDemocracy reporting, in late April 2026 it was announced that Google had signed a contract with the US Department of War—the official name is the Department of Defense—allowing the Pentagon to use Google AI for "any lawful government purpose," including "mission planning and weapons targeting" [1][2]. The contract details have not been independently confirmed by public documents. Liguori and over 600 colleagues petitioned CEO Sundar Pichai to oppose the partnership; the petition received no response [1][2].

Liguori's resignation letter states: "I no longer believe that our technology will have a net positive impact on society... it's not this type of general-purpose AI we are racing to develop" [1][2].

NDAs and Google's concession

Multiple UK-based Google employees who left on ethical grounds were bound by lifetime non-disparagement agreements. Liguori breached hers to publish her resignation letter. This week, Google told UK staff it would not enforce the post-termination non-disparagement clause. A spokesperson told The Washington Post the company had "updated our terms for new hires, which included removing the post-termination aspect of non-disparagement clauses" [1][2].

The change follows organizing by DeepMind members of United Tech and Allied Workers (UTAW) with No Tech for Apartheid and whistleblowing nonprofit Psst, and coincides with the UK's Employment Rights Act 2025 taking effect [1][2].

What remains unclear

Google has not responded to openDemocracy's request for comment at time of publication [1][2]. The October 2025 press release describes only unclassified deployment in phase one [5]. Whether classified deployment has begun, is planned, or was aspirational in internal announcements is not confirmed by public documents.