Google Quietly Removes AI Weapons and Surveillance Pledge From Principles

This article was prepared from the source material provided, which includes reporting by Bloomberg and The Intercept, a quote from Margaret Mitchell, a Google statement, and a 2019 quote from Shoshana Zuboff.

Google has quietly removed commitments not to use its artificial intelligence for weapons or surveillance from its published AI principles, a change that drew immediate criticism from a former leader of its ethical AI team.

The original 2018 statement pledged that Google would not design or deploy AI for uses "likely to cause overall harm," explicitly including weapons and surveillance technology. Those commitments no longer appear on the company's AI principles page, according to Bloomberg, which first reported the change.

Margaret Mitchell, former head of Google's ethical AI team, told Bloomberg that the removal erases years of internal advocacy. "Having that removed is erasing the work that so many people in the ethical AI space and the activist space as well had done at Google," she said. "More problematically it means Google will probably now work on deploying technology directly that can kill people."

An Industry Pattern

Google is not the first major AI developer to step back from such restrictions. Last summer, OpenAI deleted its own pledge not to use AI for "military and warfare," as reported by The Intercept at the time.

In a statement, Google said that "companies, governments, and organizations… should work together to create AI that protects people, promotes global growth, and supports national security." The company has not announced specific military products.

Google's Longer History With Data

The shift comes amid heavy investment in AI across the tech industry, with companies under pressure to generate returns for investors. Tech critic Shoshana Zuboff, writing in 2019, described Google as "the pioneer, discoverer, elaborator, experimenter, lead practitioner, role model, and diffusion hub of surveillance capitalism."

Google began exploring the value of personal browsing data — sometimes called "digital exhaust" — in the early 2000s, recognizing it contained predictive information about users. Pressured by the Dot-com collapse, the company built that data into the dominant tracking and advertising apparatus of the modern web, spinning off hyper-targeted ads, predictive algorithms, personal assistants, and smart homes.

What remains unclear is how the removal of the weapons and surveillance commitments will translate into actual products or contracts. Google has not detailed any new military or surveillance work, and the practical consequences of the change have yet to be demonstrated.