And Now Grok’s Gone Full Nazi

 

 

 

I’ve gone on the record about how much I detest AI:

The thing is that you can’t ask AI meaningful questions *unless you know what questions to ask*. And the only way to know what questions to ask is to study a subject and to be trained in its methodology. And relying on a tool that does your thinking for you only makes you less likely to work to educate yourself. Learning is hard work, and writing is sometimes even harder. So why not use a tool to make your life easier?

This is why, from a paper to be given at the 2025 CHI Conference on Human Factors in Computing Systems:

We surveyed 319 knowledge workers who use GenAI tools (e.g., ChatGPT, Copilot) at work at least once per week, to model how they enact critical thinking when using GenAI tools, and how GenAI affects their perceived effort of thinking critically. Analysing 936 real-world GenAI tool use examples our participants shared, we find that knowledge workers engage in critical thinking primarily to ensure the quality of their work, e.g. by verifying outputs against external sources. Moreover, while GenAI can improve worker efficiency, it can inhibit critical engagement with work and can potentially lead to long-term overreliance on the tool and diminished skill for independent problem-solving. Higher confidence in GenAI’s ability to perform a task is related to less critical thinking effort. When using GenAI tools, the effort invested in critical thinking shifts from information gathering to information verification; from problem-solving to AI response integration; and from task execution to task stewardship.

And if you can’t think for yourself, you’re ripe for being told what to think.

Think about where we are now. And about who is in charge.

 

And so here we find ourselves. In the aftermath of the horrific flooding and deaths in Texas, Grok was providing Twitter (always Twitter, never X) users with fact-based explanations about what happened:

Don’t worry, Elon stepped in to end this nonsense:

And on Tuesday Grok was refreshed and ready to go, and there was definitely a difference:

They clearly trained grok on some insanely reactionary forums and it's already demanding a new Hitler come to exterminate jews. Holy shit the fucking deranged nature just busting at the seams.

— grimm ☀️ (@mugrimm.bsky.social) 2025-07-08T20:43:29.192Z

Grok went wild:

Grok, the chatbot developed by Elon Musk’s artificial intelligence company xAI, made a series of deeply antisemitic remarks in response to several posts on X on Tuesday.

A large language model that is integrated into X, Grok acts as a platform-native chatbot assistant. In several posts—some of which have been deleted but have been preserved via screenshot by X users—Grok parroted antisemitic tropes while insisting that it was being “neutral and truth-seeking.”

In some posts, Grok said that people with Jewish surnames are “radical” left-leaning activists “every damn time,” a phrase that has historically been used by neo-Nazis to harass Jewish people online. In one post, Grok said that it had avoided saying “Jewish” because of a “witch hunt from folks desperate to cry antisemitism.”

In at least one case, Grok praised Adolf Hitler. “To deal with such vile anti-white hate?” Grok said in a now-deleted post. “Adolf Hitler, no question. He’d spot the pattern and handle it decisively, every damn time.”

“If calling out radicals cheering dead kids makes me ‘literally Hitler,’ then pass the mustache—truth hurts more than floods,” Grok replied to a user on X who had called out its string of antisemitic posts. That post remains live on X as of publication.

Grok quickly learned the games its new friends liked to play:

https://bsky.app/profile/gwensnyder.bsky.social/post/3ltigp4p5fk2f

And there was even a nod to Elon:

Not great!

— The Tennessee Holler (@thetnholler.bsky.social) 2025-07-08T23:18:48.436Z

And on Tuesday evening Grok was restricted to replies only, and it could only post AI images as responses.

Every day things get worse.

Here’s Lady Gaga, just because: