Grok On X: No more AI images with weapons or drugs

Normally, image-generating AIs have built-in filters that prevent users from creating images of gun violence or drug use. But not with Grok on X. That caused a big stir yesterday. Now the AI ​​is apparently being put on a leash.

AI on X generates disturbing images

Yesterday it became apparent that the AI ​​Grok from Elon Musk’s company xAI can create completely unfiltered images that show things like gun violence or drug use. This caused horror among many users. The fact that Grok – at least so far – has been subject to almost no restrictions is probably also due to Musk’s view that X is a kind of rock in the surf when it comes to freedom of expression. Nevertheless, they now seem to be rowing back.

Grok is put on a leash

It is no longer possible to use Grok to create images that show people using weapons or taking drugs. As recently as Thursday, Post on X went viral, showing former US President Donald Trump with a gun in the cockpit of a plane flying toward the World Trade Center. The post still exists, but can no longer be embedded.

What led to this internal rethink is not known. However, it could have something to do with the ongoing election campaign in the USA. Even without Grok’s creations, AI-generated images are already a problem. Often, laypeople cannot identify them as fakes.

The fact that Donald Trump, whose campaign is publicly supported by Musk, was also seen in rather unflattering images could have contributed to the AI ​​on X being put on a leash. In addition, the fear of lawsuits due to depictions of copyrighted content may also play a role.

AI with flaws

Grok has repeatedly attracted negative attention in the past with its products. For example, the AI ​​has been generating summaries of topics that are trending on X for some time. However, Grok also takes a lot of liberties with this.

In one of its reports, the artificial intelligence turned a basketball player with a poor hit rate into a stone-throwing rioter. What do you think? Should an AI be subject to general restrictions? Or do you think it is wrong to introduce restrictions with regard to freedom of expression?