A new report from the information says Alphabet and the Pentagon are discussing an AI deal.

This is the military looks to revamp its ties to AI after that, of course, significant fallout it had with Anthropic.

McKenzie Segalos is looking into all of that in today's tech check, Mac.

Hey, David. So those deal talks between Google and the Pentagon would let the military use Gemini in

classified settings. According to the information, Google has proposed contract language barring uses

like mass domestic surveillance and fully autonomous weapons without human oversight. But the broader

agreement, at least according to this report, would still allow use for all lawful purposes, the same

legal framework that Anthropic refused to accept back in February. And that's what's so

significant. It's not just that it would dramatically expand Google's defense business. It's

that it comes only months after Anthropic's own talks with the Pentagon collapsed over many of those

same issues. Now, there were already

signs that Google was moving to capitalize on that fallout. Google Cloud Chief Thomas Currian

reportedly pitched the Pentagon on a broader partnership the very day Anthropics negotiations

fell apart. And within weeks, Google was offering AI agents in unclassified government settings,

and they weren't alone in spotting that

opening. OpenAI signed its own Pentagon deal for classified workloads in the immediate aftermath

of Anthropic Talks imploding. XAI followed soon after. But Gemini was really the big

question mark, because Google had the most fraught history with the military of any of them. This is the company

that let its Project Maven drone contract with the Defense Department expire after employees

revolted internally. Its motto used to be, don't be evil, and now it's in talks to put its

AI into classified Pentagon systems. It's pretty much a 180 from where they were. And the runway is there,

right? Alphabet already removed its old ban on AI for weapons and surveillance from its public

...