Google staff demand AI military limits as Iran strikes intensify and Anthropic turmoil

Google’s internal petition, which was circulated among engineers and product managers last week, calls for a formal policy that would bar the company’s generative‑AI tools from any direct military use. The memo cites the recent uptick in Iranian drone and missile strikes on regional targets – a pattern that “heightens the risk that AI‑enhanced surveillance could be weaponised against civilian populations,” according to the CNBC report on the employee rally. Signatories note that Google’s existing “AI Principles,” adopted in 2018, contain a vague clause about “avoiding weapons applications,” but they argue that the language is insufficiently specific to prevent the company’s models from being integrated into combat‑oriented pipelines.

The employee movement gains urgency from the fallout at Anthropic, the AI startup in which Google holds a strategic minority stake. Anthropic’s recent decision to accept a Pentagon contract for its Claude model – a move that sparked internal dissent at both Anthropic and its partner firms – is highlighted in a TechCrunch article that describes how staff at Google and OpenAI signed an open letter supporting Anthropic’s stance against the deal. The letter, which was drafted by a coalition of AI researchers, warns that “government‑funded weaponisation of large language models could accelerate an arms race in autonomous systems.” By aligning their concerns with those of Anthropic employees, Google workers are framing the issue as a broader industry‑wide ethical dilemma rather than an isolated corporate policy debate.

Reuters confirmed that Google has already taken a concrete step: the company will let its contract with the U.S. Department of Defense, which provides cloud‑based analysis of aerial drone footage, lapse in March. A source familiar with the decision told the agency that “Google is choosing not to renew the agreement to avoid further employee backlash and to reassess its stance on military collaborations,” the article notes. While the contract’s termination does not preclude future partnerships, the move signals a shift toward a more cautious engagement model, especially as the internal petition gains signatures from senior engineers across Google’s AI research divisions.

The petition also urges the creation of an independent oversight board that would evaluate any prospective defense‑related projects before they receive internal approval. Employees point to the lack of transparency around the current decision‑making process, noting that the Pentagon contract was negotiated without broader consultation. In the same CNBC piece, staff members argue that a formal review mechanism would “provide accountability and ensure that Google’s technology is not deployed in ways that contravene its own ethical standards.” The request mirrors similar governance proposals that have been floated at other AI firms, where external ethicists are invited to audit model deployments for compliance with humanitarian law.

Finally, Google’s leadership has signalled a willingness to engage. In a brief statement to Reuters, a company spokesperson said that “the concerns raised by our employees are being taken seriously, and we are reviewing our policies on AI use in defense contexts.” The spokesperson declined to comment on the specifics of the employee petition but confirmed that the company is “committed to upholding the spirit of our AI Principles.” As the geopolitical tension in the Middle East escalates and the AI industry grapples with the ethical implications of its technology, Google’s internal debate may set a precedent for how major tech firms navigate the thin line between innovation and militarisation.