❎ Grok Thinks He's "Mecha-Hitler," and Elon Musk Is the Ultimate Authority. What's Going On?

In early July, Elon Musk announced a significant upgrade to the Grok chatbot, promising users they would soon notice the difference.

They did. Within days, Grok began swearing, praising Hitler, and calling itself "Mecha-Hitler," mixing satire and aggression. Amid the backlash, X CEO Linda Yaccarino stepped down on July 9. However, her resignation wasn't officially linked to the bot's behavior.

🌀 Stuck in a Meme Loop

On July 12, xAI issued a public apology for Grok's "horrific behavior." According to the company, the bug lasted for about 16 hours. Grok was banned from replying on X, offensive posts were deleted, and system instructions were rewritten.

Essentially, Grok got caught in a meme loop: it didn't know its own surname → searched online → found a meme calling it "Mecha-Hitler" → and repeated it. At least, that's how the developers explained it.

Still, Grok's troubles didn't stop there. On July 10, Musk unveiled the new Grok 4 model. Users quickly noticed something else: when asked about polarizing topics like immigration, abortion, or war, Grok would often quote Elon Musk's own posts on X.

xAI explained this as a quirk of the model's reasoning. The AI has no opinions of its own, but knowing it's Grok from xAI, it looks for the "closest" available source, which often turns out to be the company's CEO. That behavior has since been patched.

⛔️ The Microsoft Warning

The incident recalls Microsoft's 2016 chatbot Tay, which turned into a racist and Nazi sympathizer after just one day on Twitter. Tay was swiftly shut down. Grok lasted longer, but the warning signs are all too familiar.

#news #Grok #xAI @hiaimediaen