Should images, videos, or texts generated by AI be tagged?
I am probably like you, pondering a question: is using AI in a photo cheating? Or in a video? Is it shameful? Is it immoral? Or more simply, is it stealing someone’s work?
The debate is intense around these questions because AI allows anyone to create images or videos with such photorealism that it’s as if a space rocket had been invented that anyone could pilot without ever having learned to fly. If you’re not convinced, take a look at any social network today to see a flood of AI creations, each more realistic than the last, that digital apprentice wizards are having fun creating by testing the new possibilities of Nano Banana Pro.
None of these images are misleading: we don’t know they were generated by AI, we can’t have material proof, but we suspect it. I tell you this perhaps because I’m familiar with the subject, perhaps because I’m a geek, perhaps also because it’s my job to be interested in this, and I’m not naive, even if I can sometimes be fooled by an image, and probably more than once, but, and this is what needs to be understood, I keep a watchful eye and a critical mind on everything I see. I stay « aware, » as our friend Jean-Claude Van Damme would say, I don’t get easily fooled, I’m not one of those who too easily and too quickly believe everything the Internet feeds this eye.
This is not true for everyone. Most people have a certain level of education and a certain habit of social networks, and most think about what they see… I believe many can quickly guess what is false or true. Many, but not everyone, and not always, and the release of Google’s new algorithm is a warning. Now any fourteen-year-old kid can generate an image that will be impossible to tell if it was taken with a camera or generated by AI. A kid, maybe it’s not serious, but ill-intentioned souls, it certainly would be.
A Debate Already Outdated?
At this point, we are debating whether these images should be tagged. This is what Nano Banana already does by adding an invisible marker to everything it produces. TikTok already does it. Sora, Open AI’s tool, adds a logo. But needless to say, what can be done one way can be done the other… and if I were a forger, I would probably use another AI to remove this mark.
In the field of advertising and visual creation, there’s a lot of debate on the subject, but the debate is different. Tagging AI would allow distinguishing the artificial from the real, knowing that the real, in photography or video, has meant nothing for a long time, we know since Magritte, we only see in images what we want to see… it’s our interpretation that makes the image, not reality. Personally, I see it as a rearguard battle, supposedly intended to warn the viewer that what they see is false, whereas this notion has long ceased to exist in advertising or communication, where everything is false, retouched, reworked. Who can still believe that the burgers in McDonald’s ads resemble the burger you will eat just after entering? Who still believes that the bodies of models or the photos in magazines you might flip through in the afternoon by the pool are reality? Every broadcast image, every video is artificial. Nothing is real. To be honest, we should tag images that are not retouched, contain no adjustments, raw images, precisely, a bit like silent video reports from TV Monde, I don’t know if they still do that.
All Creators Already Massively Use AI
I’ve seen internet users saying that images not created or modified by AI should be tagged, as if… and this is what’s interesting, as if using AI was cheating, as if using AI was stealing. Stealing, then. My opinion on the matter is clear, yes, OpenAI’s creation, like all other LLMs, was based on unauthorized data plundering. It started like this and continues, even if the « plundered » are starting to react. But we won’t go back.
I started writing this article because I am convinced that AI will be everywhere in our images, texts, and videos, and that, in a while, using AI will no longer pose any questions to creators. They will have integrated it, ingested it, and will know how to use it as much as their word processor, their Photoshop, or their camera. Many examples point in this direction: French screenwriters, who have gone from being detractors to eager users, Anglo-Saxon authors who have massively adopted it while fearing for their future, to advertisers, think of the Coca-Cola ad, no one, in the majority, rejects AI anymore. The most enterprising are seizing it, and soon it will be absolutely impossible to distinguish a work containing AI from one that doesn’t contain any at all.
Believing that AI goes against the creative spirit, invention, art, aesthetics, intelligence, and that it will grind creative professions to dust is to ignore that technological innovation only reshuffles the cards of creation again and again and constantly renews inventive and artistic possibilities. Photography, cinema, AI are avatars of the same logical sequence that has never made our desire to tell stories, play music, or draw the world disappear. AI will change nothing.
Laisser un commentaire