I know — you’ve probably heard that using ChatGPT (or any bot) to learn is totally fine. But yeah… scratch that.

Because apparently, even Kim Kardashian — reality TV royalty, business mogul, and future lawyer — just revealed her “AI study buddy” straight-up made her fail her law exams.

In a Vanity Fair interview, Kim confessed she’s been relying on ChatGPT for legal questions. Like she'd literally snap a pic of her test materials, toss it into the bot, and wait for the magic.

Except the magic wasn’t real — it was hallucinated. 

And the results? Totally wrong. She says the AI’s bad advice made her fail tests.

So, for the millionth time, here’s what’s actually happening:

ChatGPT doesn’t know facts. It predicts text based on patterns — and when it doesn’t have real data, it just makes things up (confidently).

It’s like that one friend who says everything with a straight face, even when they’re guessing.

And get this: these AI hallucinations are a massive deal.

  • Lawyers have already been hit with sanctions for filing briefs citing fake cases generated by AI.

  • Students have failed assignments because their “AI tutor” lied to them.

  • And doctors? They’ve flagged ChatGPT-style systems for giving dangerously inaccurate medical advice.

Kim’s experience — plus the tragic AI cases we covered earlier, basically slaps a glossy celebrity filter on a way bigger issue. Which is: 

We’re trusting machines that sound smart but don’t actually understand anything.

To be fair, we’ve also seen some incredible wins — people using ChatGPT or Claude to challenge bogus medical bills, draft small-claims suits, handle complex legal paperwork and actually win cases.

But keep in mind — that’s the one-in-a-hundred scenario where it works beautifully. For the other ninety-nine? Proceed with caution.

The moral: AI can be helpful, but it’s not holy. Always double-check. Cross-verify. And remember — confidence doesn’t always equal correctness.

So yeah… maybe don’t fully trust ChatGPT as your law school study buddy. 😉

Reply

or to participate

More From The Automated

No posts found