Google DeepMind Staff Quit Over $1.2 Billion AI Military Contracts
Updated
Updated · bbc.co.uk · Oct 21
Google DeepMind Staff Quit Over $1.2 Billion AI Military Contracts
3 articles · Updated · bbc.co.uk · Oct 21
Summary
Several Google DeepMind employees have resigned or are preparing to leave over the company’s expanding military AI work, with researcher Sofia Liguori quitting this week and Andreas Kirsch leaving earlier this month.
Their objections center on contracts with the Pentagon and Israel, including a deal allowing Google AI to be used for classified work and a separate $1.2 billion Project Nimbus contract with Amazon for Israel.
Departing and current staff told the BBC they doubt Google has meaningful oversight over how militaries use its models and fear retaliation for speaking publicly because employment terms restrict criticism.
Google said it reviews public-sector contracts under its AI principles and acceptable-use policies, while critics argue current models are not robust enough for warfare and can outpace meaningful human review.
The dispute reflects a wider push by the Trump administration to make the US an “AI-first warfighting force,” as major tech firms sign military AI deals and concerns grow over battlefield use.
Are tech giants secretly writing the very military regulations meant to restrict their own highly lucrative defense contracts?
If the engineers building military AI refuse to trust it, what terrifying flaw are they seeing that the Pentagon is ignoring?
With defense oversight severely understaffed, who is actually controlling the autonomous systems currently analyzing targets on the modern battlefield?