Google Director Resigns Over Pentagon AI Deal, Citing Principles

René Mayrhofer, a director for Android platform security at Google, resigned over the company's agreement to provide its AI models to the US Department of Defense for classified work [1]. His departure highlights internal dissent following Google's updated AI principles in February 2025, which removed prior restrictions against using AI for weapons or surveillance tools [1, 3].

Internal Backlash Over Military AI

Mayrhofer, who joined Google in 2017, stated his decision became "unavoidable" after the company signed a deal allowing the Pentagon to use its AI for "any lawful purpose," including military operations and intelligence gathering [1]. He criticized Google management for "quietly abandoned the goals to become carbon-neutral because of the AI model energy usage" and for signing deals with the "US Ministry of War" [1]. Mayrhofer's resignation note, titled "Google Management Has Lost Its Moral Compass," also referenced Google's 2018 AI principles, which pledged against developing weapons or surveillance tools, a stance that was removed in a February 2025 update [1].

Google announced in late April that it had signed a deal with the Department of Defense to provide its AI technology for classified work [1]. This agreement follows an open letter from over 600 Google employees urging the company to avoid making AI systems available for classified workloads due to concerns about "unethical and dangerous uses," including lethal autonomous weapons and mass surveillance [3]. Google's spokesperson stated the company is "proud" to be part of a consortium providing AI services "in support of national security" and remains committed to the consensus that AI should not be used for domestic mass surveillance or autonomous weaponry without human oversight [1]. The company also noted that providing API access to commercial models under standard practices is a "responsible approach" to supporting national security [3].

Shifting AI Principles and Transparency

The tension within Google echoes a 2018 internal revolt over Project Maven, a Pentagon program involving AI analysis of drone footage, which Google ultimately chose not to renew [3]. However, the company's stance on military and national-security AI has shifted since then [3]. In a February blog post updating Google's AI principles, Google DeepMind CEO Demis Hassabis and senior vice president James Manyika wrote that "democracies should lead in AI development" and that companies and governments should collaborate to build AI that "protects people, promotes global growth and supports national security" [3].

Mayrhofer's concerns extend to a perceived lack of transparency, noting that decisions are "just being decided by top-level management" without internal debate [1]. He also expressed concern that the Pentagon deal, allowing AI for "any lawful purpose," could include mass surveillance of EU citizens [1]. Google's agreement includes language stating its AI system is not intended for domestic mass surveillance or autonomous weapons without human oversight, but also notes Google does not have the right to control or veto lawful government operational decisions [3].

What to watch

• The broader rollout of Google's AI services for national security, and any further employee responses or resignations [1, 3].
• Updates to Google's official documentation regarding its AI principles and the specific applications of its AI models in classified government work [1, 3].
• The development of additional metrics for Google's generative AI report, which is currently in limited testing, beyond impression counts [2].

The ongoing internal debate at Google over its AI principles and military contracts highlights the ethical complexities and transparency challenges as AI technology becomes increasingly integrated into national security applications.