Matt Wolfe breaks down Google's new Pentagon contract allowing its AI for any lawful government use, a stark reversal from DeepMind's 2014 ban on military applications. Over 600 Google employees have publicly opposed the deal. Wolfe argues AI companies face an impossible choice: refuse defense work and get sidelined, or accept it and face backlash.
To be fair to Google, I do think they're in a kind of no-win situation. There was recently a ton of drama around Anthropic refusing to work with the Pentagon. Well, Google just signed a deal with the Pentagon, and their agreement basically says you can use it for any lawful government purpose. They do go on to say, we remain committed to the consensus that AI should not be used for domestic mass surveillance or autonomous weaponry without appropriate human oversight. There's no like binding agreement. We're just saying we don't think you should. This is especially interesting because back when Google bought DeepMind back in 2014, there was actually an agreement that they wouldn't use it for these purposes. This understanding was framed as a core condition of the deal and part of DeepMind's mission around ethical and responsible AI. They either say no to a deal with the Pentagon, but then they risk being sort of ousted like Anthropic was. These AI companies are also needing legislation to sort of go in their favor to continue pursuing AI advancement. But then on the other hand, you've got 600+ employees over at Google basically begging Google not to do this. And then you have the optics. So Google's options were bad on both sides.
AI News: Dots, GPT-6.1 Sol, Sonnet 5.5, Gemini 4, and everything you need to know
