Hello, hello. I'm Brittany Luce and you're listening to It's Been a Minute from NPR, a show about what's going on in culture and why it doesn't happen by accident.

Earlier this summer, Elon Musk's AI chatbot, Grok, had a pretty embarrassing malfunction.

The Mecca Hitler incident.

So as background here, the way that modern AIs are trained, modern large language models, which are the chatbots that people think of

when they talk about AI, is they are fed enormous amounts of texts, all of the books we have

recorded, all of the texts of the internet, Reddit, you know, everything like that.

Out of this comes a sort of distinctive political perspective, you know, a little bit politically liberal.

That is staff writer for the argument, Kelsey Piper.

Elon Musk didn't like this. He found it kind of embarrassing that Grock, the AI, that his team had

kept disagreeing with him on Twitter. And so he started this work to train GROC to be, you know, he

framed it as more truth-seeking. This didn't go well. When they released the new GROC, people

realized very quickly they could provoke the new GROC into, you know, expressing not just right-wing

opinions, but like horrific, like, white supremacist

reactionary opinions.

At a certain point, GROC started referring to itself as Mecca Hitler, a reference to the

video game, Wolfenstein.

So this is really bad.

And XAI did apologize and take that GROC down and spend some time trying to rework it and get a more moderate grok.

We reached out to X for comment on the situation and they directed us to a post that states, quote,

we spotted a couple of issues with Grockfort recently that we immediately investigated and

mitigated. One was that if you ask it, what is your surname? It doesn't have one, so it searches

the internet leading to undesirable results, such as when it searches, picked up a viral

...