Elon Musk and Giorgia Meloni together at the cinema: what Grok is and what risks it entails
Elon Musk and Giorgia Meloni are sitting in a cinema. They look into the room, the screen behind them seems off. An image that seems to confirm a rumor that, particularly in the United States, returns every time the Premier and Tesla’s number one meet: the one that would like them to be involved in some kind of romantic relationship.
The rumors are true, but the image you saw at the beginning of this article is less true. It is in fact a photo generated with the new version of Grok, the artificial intelligence of xAI, which has been available for free on X for about a week, what until some time ago was called Twitter.
Nothing new, strictly speaking. It is a chatbot very similar to ChatGPT: you write a request and the system responds, with a text or by generating an image. The point is that right from the start Grok – until now only available to paid users of X – had a specific feature.
When Elon Musk met with Italy’s Prime Minister, Giorgia Meloni. 🇺🇸🇮🇹
— DogeDesigner (@cb_doge) September 4, 2024
Grok and realistic images
Unlike other artificial intelligences, it does not have many limitations, neither when writing nor when producing images. And therefore it is able, to name a few, to reproduce faces of more or less famous characters or produce images inspired by copyrighted products (such as Pixar films, or the Simpsons). And to do it very well, particularly in this latest version.
In fact, since it was made available for free, when browsing X it’s difficult not to come across images generated by Grok. Photos of politicians, footballers, coaches: all portrayed in particular, often funny situations. Lorenzo Pregliasco, founder of YouTrend, published on X a series of images of Italian politicians on Vespas, on the streets of Rome, from Meloni to Mattarella. At the end of the post, he asks: “The resemblance is impressive: how difficult will it become to distinguish true and false?”.
Grok, the artificial intelligence system from xAI (Musk), now also generates images with Italian political leaders (until recently it was limited to a few global figures).
The resemblance is striking: how difficult will it become to distinguish true and false? pic.twitter.com/ZACkI9ZwgV
— Lorenzo Pregliasco (@lorepregliasco) December 12, 2024
Artificial intelligence, truth and memes
It is a risk, that of the invasion of synthetic images that make it difficult to distinguish the true from the false, which has been talked about for some time. It’s the danger, in other words, that people will believe that AI-generated photos might be real. Or that, on the contrary, realistic photos can be mistaken for images generated by artificial intelligence. In short, a sort of chaos, in which it becomes increasingly complex to understand who and what to trust.
Francesco D’Isa, who has been studying the relationship between images and reality for some time, has often written about how the situation is a little more complex. In a piece published some time ago on Domus Web he wrote: “The strength of an image lies in its ability to tell a story, but only if it is supported by a credible context.”
It’s the context that makes the difference. Or rather, the story that that image, realistic or not, is trying to tell. And therefore, in Grok’s case, he explains to me: “The real problem is considering videos and images incontrovertible sources of truth: from the very moment you choose what to frame, photography is a lie. Now, falsification techniques have improved, but the problem is not technological, because misinformation passes above all through context and source. Think about Trump and the Haitian immigrants who eat dogs and cats: if your uncle had said it at Christmas dinner no one would have believed it; Unfortunately someone did it because it was the future President of the United States.”
“Culturally or instinctively, we tend to attribute to images a function of proof, of testimony – Giulio Armeni, one of the most well-known Italian memers with the profile @filosofia_coatta, confirms to me -. The point is that we are living in a historical moment in which we realize that images are no longer a guarantee of truth and objectivity. The meme and the images generated by artificial intelligence explore precisely this restlessness, this sense of instability that characterizes our time.”
In this sense, images generated by generative AI are more than anything another type of meme, a way of bending reality. “The meme, unlike a drawing, uses material already consolidated in the collective imagination. Recycle images, concepts, sentences; he never creates from nothing. It is an art of manipulation and recombination of already codified material. We are living in a historical moment in which we realize that images are no longer a guarantee of truth and objectivity. The meme and the images generated by artificial intelligence explore precisely this restlessness, this sense of instability that characterizes our time.”
