Google Android Security Director Quits Over Pentagon AI Deal

René Mayrhofer, Google’s director of Android platform security for nearly a decade, has resigned over a Pentagon deal allowing classified military use of the company’s AI models. In a May 18 farewell note he circulated to colleagues and then published online, Mayrhofer titled his message “Google Management Has Lost Its Moral Compass” and accused leadership of secretly abandoning the company’s own AI principles. His exit is a high-profile resignation from a Google workforce that has been protesting the deal for months.

The letter, obtained by Business Insider, lays out why he believes the contract is incompatible with the AI principles Google published in 2018. The deal, as The Information first reported, lets the Pentagon use Google AI on classified networks for “any lawful purpose,” a phrase that covers mission planning, intelligence gathering and weapons targeting. Google has defended the work as a contribution to “national security” and said its models are not intended for domestic mass surveillance or autonomous weapons without human oversight. Mayrhofer, who has already disconnected from AI systems that might fall under the deal, will stay at Google only through August 31, 2026, to wrap up Android security work.

The Letter That Named the Deal

Mayrhofer joined Google in 2017 as Director of Android Platform Security. Over the years that followed, he became one of the most senior internal voices defending the privacy of more than two billion Android users and a co-author of the academic Android Platform Security Model paper that codified how the operating system defends its users. In his May 18 note, he framed his departure as a moral refusal rather than a personnel decision.

I am quite sad that it had to come to this, and desperately hope Google management re-discovers its moral compass. Until then, I’ll miss y’all.

Mayrhofer wrote in the May 18 note that Google’s top management had “quietly abandoned” carbon-neutral commitments to meet the energy demands of AI training, signed the Pentagon deal without internal debate, and refused to engage employees on the change. “I am a pacifist, and have long ago decided that I will not personally work for militaries engaging in offensive warfare,” he wrote. “Proactively harming people is not something that I can or will be involved with.” The full text of the farewell note is on his personal blog.

What Google Signed in April

The Pentagon announced the deal in late April. It places Google alongside OpenAI, xAI, Microsoft, Amazon, Nvidia, Reflection AI and SpaceX on a list of frontier-AI vendors the Defense Department has agreed to use on classified networks.

The agreement lets the Pentagon use Google’s AI for “any lawful government purpose,” a phrase that, as The Information and The Guardian reported, covers mission planning, intelligence gathering and weapons targeting on classified networks. The Pentagon had pushed AI labs to accept that broad language; the company that refused, Anthropic, was designated a “supply chain risk” by the department. Google’s contract, Fortune and The Guardian noted, also obliges the company to adjust its AI safety settings and filters when the government asks.

The dollar scale is meaningful. The Pentagon’s 2025 framework agreements with each major AI lab are worth up to $200 million. Google is using the work to roll out Gemini 3.1 Pro on classified networks via the Pentagon’s GenAI.mil platform, a step first reported by Nextgov in May. The company has not disclosed the value of its specific deal.

Google’s own statement on the contract, repeated to Business Insider, said the firm is “proud” to be part of a consortium of AI labs supporting national security. The spokesperson added: “We remain committed to the private and public sector consensus that AI should not be used for domestic mass surveillance or autonomous weaponry without appropriate human oversight.”

By the numbers

• $200 million cap on each Pentagon framework agreement with a major AI lab
• 600+ Google employees who signed the April 27 open letter to Pichai
• February 4, 2025 date Google removed its weapons and surveillance pledges from public AI principles
• August 31, 2026 Mayrhofer’s last day at Google under his current contract
• 9 years Mayrhofer’s tenure at Google, beginning with his 2017 offer

The February 2025 Rewrite

Mayrhofer’s letter is unsparing about how the contract became possible. The 2018 AI principles Google CEO Sundar Pichai published pledged that the company would not pursue “weapons or other technologies whose principal purpose or implementation is to cause or directly facilitate injury to people” and would not build “technologies that gather or use information for surveillance violating internationally accepted norms.” On February 4, 2025, Google deleted those pledges from its public AI principles, a quiet change the human rights response to the 2025 ethics reversal condemned as a “dangerous precedent.”

Google’s decision to reverse its ban on AI weapons enables the company to sell products that power technologies including mass surveillance, drones developed for semi-automated signature strikes, and target generation software that is designed to speed up the decision to kill.

Matt Mahmoudi, a researcher and adviser on AI and human rights at Amnesty International, called the move “a shame” and warned that “AI-powered technologies could fuel surveillance and lethal killing systems at a vast scale.” Mayrhofer links the rewrite directly to the Pentagon contract: a company that has dropped its red lines cannot credibly object to the work that follows.

Dissent From the Inside

Mayrhofer is the most senior defector yet, but the rebellion inside Google is not his alone. The workforce has been pressing the company to drop or rewrite the Pentagon deal since the spring.

On April 27, more than 600 Google employees signed an open letter to Pichai asking him to refuse to make the company’s AI systems available for classified workloads. The letter read: “We feel that our proximity to this technology creates a responsibility to highlight and prevent its most unethical and dangerous uses.” Two days later, Andreas Kirsch, a research scientist at Google DeepMind, told Business Insider he was “incredibly ashamed” of the contract and described it as a “worst-case version of events.”

The protest has spread beyond the letter. Google DeepMind researchers in the United Kingdom voted in early May to unionize over military AI contracts, the first such move at a frontier AI lab, and have threatened “research strikes.” Alex Turner, a research scientist at Google DeepMind, said on X that he had spent two months trying to prevent the deal, calling it “shameful.” In a separate letter, more than 100 Google AI employees wrote to chief scientist Jeff Dean opposing Gemini’s use for U.S. surveillance, the New York Times reported.

Key flashpoints in the internal revolt

• February 2026: 100+ Google AI employees write to chief scientist Jeff Dean opposing Gemini’s use for U.S. surveillance.
• April 27, 2026: 600+ Google employees sign an open letter to Pichai asking him to refuse the Pentagon contract.
• April 29, 2026: DeepMind researcher Andreas Kirsch publicly says he is “incredibly ashamed” of the contract.
• Early May 2026: UK-based DeepMind workers vote to unionize, the first such move at a frontier AI lab.
• May 18, 2026: Mayrhofer publishes his farewell note titled “Google Management Has Lost Its Moral Compass.”

Why the Lever No Longer Pulls

Eight years ago, internal protest worked. In 2018, thousands of Google employees signed a letter objecting to the company’s work on Project Maven, a Pentagon contract that used Google’s AI to analyze drone surveillance footage, and Google declined to renew it. The Maven episode is the precedent the company now points to when it insists its red lines have shifted.

The Maven playbook is harder to run in 2026. Laura Nolan, a former Google engineer who resigned over Maven and spoke to Fortune, said the company has spent the years since cracking down on internal organizing. She pointed to the decommissioning of internal mailing lists and Google’s internal social network. The leverage workers once had is also weaker: cost-cutting and the threat of AI-driven automation have made collective action riskier, and Alphabet’s broader $80 billion AI infrastructure raise is reshaping what the workforce is even building. As one anonymous DeepMind researcher told Fortune, “There was a pride in doing AI for good for a very long time. Suddenly, the things I’ve pushed to improve might be used in very different ways with not enough oversight to harm people.”

OpenAI’s experience shows what internal pressure can do. Its contract used the same “all lawful purposes” language, Fortune reported, and the backlash inside the company was strong enough that CEO Sam Altman publicly apologized for the “sloppy and opportunistic” deal and said the company would renegotiate parts of it.

Google’s contract, legal experts told Fortune, appears to be the most permissive yet. The agreement obliges Google to adjust its AI safety settings and filters at the government’s request, and the language around autonomous weapons and mass surveillance imposes no enforceable obligation on the Pentagon, according to Charlie Bullock, a senior research fellow on LawAI’s U.S. Law and Policy team.

Back to the Institute in Linz

Mayrhofer’s last day at Google will be August 31, 2026. He is a tenured professor at Johannes Kepler University Linz in Austria, where he heads the Institute for Networks and Security.

He has already stopped working on any AI systems that could fall under the Pentagon deal. In his note, he said he plans to keep working on end-to-end encrypted communication, privacy-preserving digital identity, embedded systems security and operating-system and supply-chain security, all of which touch the open-source AOSP version of Android. The work is, in effect, the same job he was doing inside Google, minus the company.

The split is also a personal one. As an Austrian academic and a European citizen, Mayrhofer wrote, he believes the U.S. government has “become hostile” to educational institutions and that “any lawful purpose” under the Pentagon deal could include mass surveillance of EU citizens. “This deal implies that Google (AI) products will likely be used directly against me and mine,” he wrote. “In this recent environment, I don’t see how I could not resign.” Mayrhofer’s resignation lands alongside a Senate inquiry into Pentagon contract awards and a UK-based DeepMind unionization drive, both of which put Google’s choice under fresh public scrutiny. The company has not commented on the resignation.

Frequently Asked Questions

Who is René Mayrhofer?

René Mayrhofer joined Google in 2017 as Director of Android Platform Security, leading the team from Mountain View, California until fall 2019 before relocating to Austria and shifting to a part-time, strategic role. He is also a tenured professor and head of the Institute for Networks and Security at Johannes Kepler University Linz, where his research focuses on mobile security.

Why did he resign from Google?

In a May 18, 2026 farewell note titled “Google Management Has Lost Its Moral Compass,” Mayrhofer said the decision became “unavoidable” once Google signed a Pentagon deal allowing classified military use of its AI for “any lawful purpose.” He cited his pacifism, his position as an EU academic, and his belief that the U.S. government had “become hostile” to European institutions as reasons he could not continue.

What is the Pentagon AI deal Google signed?

The agreement, announced in late April 2026, is part of a Pentagon framework under which each major AI lab can receive contracts worth up to $200 million. It places Google alongside OpenAI, xAI, Microsoft, Amazon, Nvidia, Reflection AI and SpaceX. Anthropic is the only frontier AI lab that has refused the same terms and was designated a “supply chain risk” by the Pentagon.

When did Google change its AI ethics principles?

On February 4, 2025, Google removed from its public AI principles a pledge not to pursue weapons or surveillance tools that violate internationally accepted norms. The 2018 principles had been published by CEO Sundar Pichai; the 2025 update was defended by AI chief Demis Hassabis on national-security grounds and was condemned by Amnesty International as a “dangerous precedent.”

How have other Google employees reacted to the deal?

More than 600 Google workers signed an April 27, 2026 open letter to Pichai asking him to refuse the contract. DeepMind researchers publicly said they were “incredibly ashamed” of the work. In early May, UK-based DeepMind workers voted to unionize, the first such move at a frontier AI lab, and threatened “research strikes.”