Google's Chief Security Officer issues a somber statement: management has "lost moral standards," abandoned the carbon neutrality commitment, and signed a classified AI contract with the U.S. Department of War

Having served as the Director of Android Platform Security at Google for nine years, René Mayrhofer announced he will leave at the end of August 2026. In a public blog post, he directly called out: Google executives have quietly abandoned their carbon neutrality commitments and signed a secret AI contract with the U.S. "Department of War," all without internal discussion.
(Background: Breaking news — Pentagon emergency lockdown after "hazardous material incident"! Military personnel in chemical protective gear deployed)
(Additional context: Why AI hasn't caused large-scale unemployment among software engineers? Latest research: Humans remain irreplaceable in judgment and accountability)

Table of Contents

Toggle

• What did he do, and what did he see?
• What does "any lawful purpose" mean
• 600 signatures, and the image from 2018

In 2017, René Mayrhofer took on the role of Director of Android Security. At that time, Android had just surpassed 2 billion users, and the open-source-first culture was inspiring. Google also made a public commitment: the AI principles released in 2018 explicitly listed "applications we will not pursue," including weapon technology, surveillance that violates international law, and any system primarily aimed at harming people.

These principles are not just corporate documents but a public contract between Google and society at large. "Do no evil" evolved from a slogan into a written pledge, implying that employees' ethical judgments have a voice within the system.

In 2019, Google declined to renew Project Maven, a Pentagon project that used AI to analyze drone footage and assist military target detection. About 4,000 employees signed a petition against it, and at least 12 employees left as a result. Those who chose to leave became a moral reference point: the collective voice of employees had genuinely influenced the company's direction. It was an era when Google would retreat under employee pressure.

But by 2026, Mayrhofer recently wrote openly on his personal blog: he has chosen to leave.

What did he do, and what did he see?

Mayrhofer’s technical achievements within the Android security team are outstanding. He led the initiative to make device encryption the default in Android 10; launched end-to-end encrypted Android backup mechanisms; promoted Insider Attack Resistance, ARM MTE memory tagging technology, and privacy-first digital certificate architectures. He describes his principles in the blog:

"Make things so secure that even we can't break them, whether the device costs $1,000 or $100, whether the user is a celebrity or a refugee."

This is an ethical stance that treats security technology equally, protecting objects regardless of identity or wealth. It is also the core reason he was able to sustain nine years at Google.

However, over these nine years, Google's context has quietly shifted. He points out two things: first, Google management has secretly abandoned its carbon neutrality goal; the enormous energy consumption of AI models has increased the company's carbon emissions by 50% over the past five years, and the commitment has vanished without any public discussion. Second, and a more direct reason for his departure: Google executives are signing secret AI contracts with the U.S. "Department of War."

He is within the management chain. But he was completely unaware of this through internal channels. This is not just personal grievance but a deeper signal: such decisions are made within a closed loop that excludes senior technical leadership. When an engineer in the management chain knows nothing about major contracts that determine the company's ethical direction, the collapse of governance transparency becomes no longer abstract.

And what he plans to continue working on after leaving—end-to-end encryption, privacy-first digital identity, and OS security—is essentially a continuation of his work at Google. He has not abandoned these technologies; he has abandoned an organization that has already changed course.

What does "any lawful purpose" mean

At the end of April this year, The Guardian first disclosed that Google had signed a secret AI contract with the U.S. Department of Defense, authorizing the Pentagon to use Google AI models for "any lawful government purpose," as an addendum to existing contracts. Companies like OpenAI, xAI, Nvidia, Microsoft, and Amazon have signed similar agreements.

"Any lawful purpose" is a seemingly neutral legal phrase. Simply put, as long as the U.S. government deems an action lawful, Google's AI technology can be called upon, with no upper limit.

Mayrhofer’s concern is: the current U.S. government has renamed the Department of Defense as the "Department of War," with the slogan "Maximum destructive power, not lukewarm legality." This slogan clearly positions legal compliance as a baseline rather than a boundary; if it’s legal, it can be done, and only if it’s not sufficiently destructive does it warrant reflection.

For Mayrhofer, a European scholar and pacifist, "any lawful purpose" is not abstract: the Google AI products he helped build could be used for surveillance targeting EU citizens or supporting military applications he considers to violate international law. This is not hypothetical ethical debate but a real situation affecting his community: his family, his long-serving open-source community, could all fall within this framework.

The existence of this contract is not a technical issue, nor a business issue. It is a direction statement.

600 signatures, and the image from 2018

There were voices within the company. Over 600 Google DeepMind and Cloud employees earlier this year signed an open letter to CEO Sundar Pichai, demanding rejection of secret military AI work, but ultimately, they could not influence senior management decisions.

Structurally, this contrast reveals a clear shift in power: in 2018, 4,000 employees’ signatures forced the company to withdraw; in 2026, 600 signatures, and management not only refused to retreat but doubled down.

The collective voice of employees has diminished, while management’s resolve has grown stronger. As every major company in the industry has signed similar agreements, the moral pressure exerted on any single company is diluted within the industry’s collective action, and no one bears the cost of refusal alone.

Mayrhofer’s contract notice period ends on August 31, 2026, but he states he will immediately cease any AI work that could fall under this Department of War contract. After leaving Google, he will continue to focus on end-to-end encryption, privacy-first digital identity, and OS security.

An engineer who builds security technology based on the principle of "unbreakable even by ourselves" ultimately cannot find a foothold in a company that is changing course.