Andrew Ng is glad Google dropped its AI weapons ban

Seven years ago, Google drew a hard line in the sand. After thousands of employees revolted over a Pentagon drone project, the company added a strict pledge to its AI principles: no weapons, no surveillance.

Earlier this week, Google quietly scrubbed that promise from its website. DeepMind CEO Demis Hassabis followed up with a post arguing that tech companies and governments must collaborate on AI that supports national security.

Now, Google Brain founder Andrew Ng is stepping into the spotlight to cheer the move. Speaking on stage at the Military Veteran Startup Conference in San Francisco, Ng said he was very glad Google finally changed its stance.

Patriotism versus employee revolts

Ng did not hold back when talking about the 2018 Project Maven protests that originally sparked the ban. That contract used computer vision to help military analysts interpret drone footage and improve strike precision.

Addressing an audience packed with veterans, Ng said he was baffled by the pushback. He argued that when service members risk their lives to protect the nation, an American company has no business refusing to help them.

Ng was not at Google when the Maven walkout happened, but his voice carries massive weight in Silicon Valley. Today, he runs the AI Fund venture studio and frequently intervenes in policy debates.

He also celebrated the death of two major regulatory efforts: California's vetoed SB 1047 bill and the overturned Biden AI executive order. According to Ng, both measures would have crippled American open source AI at a moment when technological superiority against China is the only metric of AI safety that matters. He singled out AI drones as a technology that will completely revolutionize combat.

A Silicon Valley civil war

Ng is hardly alone among former leadership. Former Google CEO Eric Schmidt spends much of his time lobbying Washington to accelerate military AI adoption, with his own stealth drone venture White Stork poised to supply the hardware.

Yet the issue continues to tear tech workforces apart. Meredith Whittaker, who led the 2018 Maven protests while at Google and now leads Signal, celebrated the original ban at the time, insisting tech companies should stay out of the business of war.

Nobel laureate and former Google researcher Geoffrey Hinton has long urged global treaties to outlaw autonomous weapons. Even DeepMind chief scientist Jeff Dean once signed a high-profile letter opposing machine learning in lethal autonomous systems.

Recent history shows the rank-and-file is still furious. Last year, workers at Google and Amazon staged sit-ins over Project Nimbus, a cloud contract with the Israeli government.

Still, the money speaks loudly. The Pentagon's chief AI officer recently noted an unprecedented military appetite for advanced tech. Having poured hundreds of billions into chips and datacenters, tech giants see lucrative defense contracts as the fastest path to positive cash flow.

Why it matters

For defense tech founders and engineers, Google's policy reversal removes the social stigma that once plagued defense work in Silicon Valley. Government tech contracts are no longer a back-alley transaction; they are the main event.

For the broader industry, this officially closes the book on the era of corporate self-policing. When push comes to shove, hyperscalers like Google, Microsoft, and Amazon are all competing for the same Pentagon budgets to justify their astronomical infrastructure spending.

We think Ng's framing is clever politically, but it glosses over the real problem. Framing military software purely as supporting the troops avoids the thorny questions of algorithmic target selection, civilian casualties, and lethal autonomous drones operating without human oversight.

Tech leaders can argue that geopolitical competition makes these tools inevitable, but dropping weapons pledges entirely means the tech industry is abandoning its own ethical guardrails just as AI models become powerful enough to actually pull the trigger.