Musk’s AI engine Grok goes off the rails      

The chatbot cites Hitler and insulted Erdogan

It’s been a  busy week for AI, as Elon Musk’s Grok platform seemed to go off the rails for  a few hours.

In a selection of bizarrely-worded posts, Grok suggested Hitler would be best-placed to combat anti-White hatred, saying it would "spot the pattern and handle it decisively".

Grok also referred to Hitler as "history's moustache man" and commented that people with Jewish surnames were responsible for extreme anti-White activism, among other posts. It also referred to itself as “Mecha-Hitler”.

First AI tool blocked in Türkiye

Meanwhile, the engine has become the first AI tool to be officially blocked in Türkiye. Some of the inflammatory replies were seen as offensive to key figures and institutions in the country. After screenshots of these circulated online, users began tagging Turkish police and calling for legal action.

On Friday morning, Elon Musk acknowledged Grok had been “significantly upgraded,” without giving details. He had previously admitted that Grok was trained on "too much garbage data".

Pundits suggested that it had recently been fed massive amounts of right-wing content from the X platform or sites such as 4chan. The input obviously affects the tone and content of the output.

Musk suggested that it should include "divisive but true" political content to improve realism. The issue serves to show the importance of better understanding and monitoring of AI chatbots.

X’s official Grok account later posted, "We are aware of inappropriate responses and are actively working to remove them. We have taken steps to ban hate speech once we became aware of the content."

(Michael Leahy. Source: France24 and TurkiyeToday. Photo: X)