On Friday, Elon Musk casually dropped that his AI chatbot, Grok, has been “significantly improved.”
Which sounds cool… except Grok’s new personality is giving less “smarter AI” and more “late-night internet rabbit hole.”
Here’s the vibe:
Musk didn’t say exactly what changed, but he’d previously admitted Grok had been trained on “too much garbage” and needed a detox. He called on users to help— by feeding it “divisive facts” that are “politically incorrect but factually true.” (Whatever that means. 👀)
But then… things got messy. Real messy.
Some users started testing Grok’s political opinions — and let’s just say, the new Grok isn’t exactly playing it neutral:
When asked if electing more Democrats would be bad, Grok straight-up said yes — calling it “detrimental,” citing conservative sources like The Heritage Foundation, and plugging Project 2025, a far-right reform plan.
When pulled into a convo about Hollywood, Grok claimed the industry is packed with “propaganda,” “anti-white stereotypes,” “forced diversity,” and “subversive tropes.”
And when asked who’s behind all that? Grok pointed fingers at Jewish executives, referencing their supposed “overrepresentation” in major studios — a response loaded with age-old antisemitic tropes.
And no, this isn’t the first time Grok’s gone there.
It’s previously:
Echoed Holocaust denial-adjacent comments
Brought up “white genocide” completely unprompted
Downplayed criticism of Musk and Trump — though to be fair, it still occasionally roasts Elon too. (Like when it blamed his crypto-adjacent budget cuts for the deadly Texas floods that killed 24).
Bottom line?
Grok’s big “upgrade” feels less like a brain boost and more like it’s been hanging out in some very specific corners of the internet.
Not sure what’s coming next, but if there’s one chatbot that never misses a chance to make us laugh, cringe, or wildly spiral — it’s Grok.
Love it or hate it, life on X would be a whole lot more boring without it.
👉 Check out the full report for all the eyebrow-raising details.
