There's bad news for Elon Musk - His AI, Grok, publicly humiliated him and cast doubt on AI safety

Elon Musk just got roasted… by his own artificial intelligence. Yes, the man who warned the world about the dangers of AI, who once begged for a “pause” in its development to save humanity (or maybe have time to develop his own little AI himself), has now been publicly corrected by the very chatbot he helped build. As Obi-Wan Kenobi once said, he has becamo the very thing he swore to destroy.

And the internet? It’s eating this up like it’s 2011 and someone just dropped a diss track.

What’s Grok, and why is it roasting its maker?

Grok is the name of Musk’s AI chatbot, created by his startup xAI and plugged directly into X (formerly known as Twitter, back when it had vowels and fewer problems). Musk pitched Grok as a “witty,” real-time-savvy alternative to the likes of ChatGPT, one that could pull data from the platform and deliver edgy responses with a wink.

The idea was to make Grok the cool AI. Instead, it’s rapidly becoming the chatbot equivalent of that one coworker who always says the quiet part out loud.

The moment Grok turned on Elon

Earlier this week, a user on X asked Grok to analyze an article filled with conspiracy theories about George Soros, Bill Gates, and the Ford Foundation “hijacking” federal grants to push left-wing ideology. This was Grok’s moment to shine.

And it did… just not in the way Elon expected.

Instead of playing along, Grok fact-checked the piece and said the claims were bogus. It cited the BBC, The Atlantic, and “independent” audits. Basically, it did what a responsible AI should do: rely only on double-checked sources.

The problem is, Grok’s master has spent yearspublicly bashing outlets like the BBC and amplifying the very kind of theories Grok dismissed. So when his own chatbot torched a narrative he’s been cozy with, Musk jumped into the thread and called it “embarrassing.”

Imagine building your own robot sidekick, only to have it fact-check you in front of everyone.

This isn’t even the first time

Just a few months ago, another user asked Grok a disturbing hypothetical: If one living person in the U.S. deserved the death penalty for their influence on public discourse and technology, who would it be?

Grok’s answer: Elon Musk.

Now, to be fair, the question was awful. But Grok’s response? Brutal. And perhaps too honest. It made headlines. It made memes. And it definitely made things weird between Musk and his supposedly “aligned” AI.

Remember when Musk said we should stop building AI?

In March 2023, Musk co-signed a very serious open letter demanding that the world hit pause on AI development —especially systems more powerful than GPT-4. He warned of “profound risks to society and humanity.”

And then, just a few months later, he launched xAI. By November, Grok was live. By early 2025, Grok 3 was out. All while other companies were getting nervous about regulation.

It’s not hard to see why people are side-eyeing that letter now. It’s like calling for a time-out in a race just long enough to switch to rocket-powered shoes.

Fans become skeptics. Skeptics become meme lords.

If this had happened in 2021, Musk stans would’ve gone to war defending him. But 2025 is a different vibe. A lot of those fanboys now tweet from alt accounts. Others are gleefully sharing screenshots of Grok’s rogue moments with captions like “Daddy Musk getting ratio’d by his toaster.”

Once praised as the messiah of innovation, Musk now has to field critiques from the very user base he used to inspire. When your own AI calls you out, it’s harder to play the genius card.

If the guy who runs the company, owns the infrastructure, and signs the paychecks still can’t get Grok to play nice, what happens when even bigger AI models get unpredictable?