The Pentagon plugged Elon’s meme bot into classified systems. We’re cooked.

A chatbot that was literally marketed as a “lol” machine reportedly got wired into classified DoD systems.

Yes, Grok. The app that will argue with you about anime and then try to sell you premium.

According to Tech Scope News, xAI’s Grok has been integrated into parts of the Defense Department’s classified workflow, and it allegedly sparked a resignation from someone who didn’t love the idea of “battlefield decisions” meeting “edgelord autocomplete.”

The DoD will say this is about “modernizing analysis,” “accelerating decision advantage,” and “keeping pace with near-peer adversaries.”

And xAI will say it has “robust safety guardrails” and “responsible AI.”

This is the real AI arms race: not Terminator, but procurement. Whoever gets embedded first becomes infrastructure, and infrastructure gets paid even when it’s wrong. Especially when it’s wrong, because then you need a “Phase II remediation contract” that costs exactly one aircraft carrier.

Meanwhile, the person resigning over guardrails is the only character in this whole movie who isn’t auditioning for a future congressional hearing titled “Why Did The Robot Suggest Airstrikes?”

The scariest part isn’t that Grok might hallucinate. It’s that the bureaucracy will treat the hallucination like a “data point,” and some colonel with a KPI will treat it like a deadline.