It is able to create images without censorship and users are already taking advantage of it

If there is one thing that Elon Musk does not lack, it is controversies; even if we limit them to those related to their artificial intelligence projects, there are enough to fill a book. Just two days ago, the Spanish Data Protection Agency, along with eight other European agencies, received complaints for using data without authorization to train their artificial intelligence. And now, this same AI is making headlines for another controversial reason.

X, the social network that was formerly called Twitter, announced the launch of the Grok-2 beta versionthe new version of its language model on which the AI ​​of the same name is based. This is a very important announcement, given that it is the response to OpenAI’s GPT-4o and Google’s Gemini 1.5 Flash, new models that already reach a very advanced level in conversational capacity with the user.

However, the big news about Grok-2 compared to its predecessor has nothing to do with what its competitors are doing, and focuses on AI image generation. On paper, this should not be such a big announcement; after all, ChatGPT and Gemini are already capable of creating images of almost anything we can imagine. But Grok-2 is different, because It’s not “censored”.

The use of this word is somewhat controversial. It would be more appropriate to say that Grok-2 has very few filters, and has no problem generating images that other AIs refuse to create; and that assumes that it is already used to create political propaganda, images of women in swimsuits and famous people in embarrassing situations, among others. As soon as Premium X users realized this, they filled the social network with fake images, each one more bizarre.

Among the most absurd examples are the images of Barack Obama goes on a cocaine binge in a mirror, or Donald Trump riding a SpaceX rocket; Recall that this same week Elon Musk organized an interview with Trump whose success is still uncertain due to the number of technical problems he encountered.

Given that it’s an election year, it’s no surprise that many of the images created feature the candidates, particularly Trump, who appears smoking in front of the American flag or firing two guns. Although perhaps the most shocking image is one that shows him next to a pregnant Kamala Harris, the Democratic nominee, suggesting he’s the father.

Initially, most AI imaging services offered this degree of freedom; but little by little, each one set up filtered lists of terms that they could not use, such as the names of famous people, criminal or sexual acts, to avoid criticism and possible lawsuits. But Elon Musk seems to bet that this freedom will encourage users to pay the 8.47 dollars per month that Premium costs in Europe.