Google Director Resigns Over Pentagon AI Deal, Calls Out 'Lost Moral Compass'

René Mayrhofer, Google's director for Android platform security, has resigned over the company's decision to provide AI technology to the U.S. Department of Defense for classified work. In a farewell letter dated May 18 — titled 'Google Management Has Lost Its Moral Compass' — Mayrhofer called his resignation 'unavoidable' after Google inked the Pentagon deal, which allows use of its AI models for tasks including military operations planning and intelligence gathering.

Mayrhofer's letter, later posted publicly on his personal blog, took aim at two major shifts: Google's quiet retreat from its carbon-neutral commitments due to AI energy demands, and its active pivot toward military contracts. He specifically criticized the deal's broad 'any lawful purpose' language, arguing the current U.S. government has already demonstrated violations of international law. As a self-described pacifist and tenured professor at Johannes Kepler University Linz in Austria, Mayrhofer said he feared the arrangement could be used to surveil EU citizens — potentially including himself and colleagues.

The resignation points to a deeper cultural rupture at Google. In 2018, the company published AI principles explicitly pledging not to develop weapons or surveillance tools. In February 2025, those restrictions were quietly removed. Mayrhofer joined Google in 2017, but said the company he signed up for no longer exists, criticizing top management for making sweeping ethical decisions without internal debate or transparency.

Mayrhofer is not alone. Andreas Kirsch, a Google DeepMind research scientist, told reporters in April he was 'incredibly ashamed' of the Pentagon deal. Google, for its part, has said it is 'proud' to participate in a consortium of AI providers supporting national security, while maintaining it opposes autonomous weapons and domestic mass surveillance without human oversight. Mayrhofer will work through his notice period until August but said he would immediately stop contributing to any AI systems covered by the Defense Department agreement.