Can you imagine two presidents of rival sports clubs doing that? On the left, Nasser Al-Khelaïfi, big boss of Paris Saint-Germain, big delighted smile. On the right, with a gentle and submissive look, his Marseille counterpart Pablo Longoria, leaning against the bust of the Qatari leader.
On September 27, three days after OM’s defeat against PSG (4-0), the image circulated widely among football fans. Many people feared being scammed. Was this not a “fake” (artificial) image, signed by artificial intelligence? In the age of AI, we are all called upon to think like journalists.
Photorealistic chimeras
We now know since the emergence in the spring of MidJourney, an algorithm capable of creating a fictitious image from online photo banks, nothing is simpler today than to give a photorealistic form to our most chimeras. crazy.
Haven’t we seen pseudo-romantic moments between Joe Biden and Donald Trump, here in an imaginary game of golf, there in a fabulous pillow fight? Nothing, or very little, in the grain of the images allowed one to infer their falsity, apart from the total absence of a serious source, and their striking implausibility.
Isn’t it the same here? There is something intuitively wrong about this hug between Nasser Al-Khelaïfi and Pablo Longoria. Anyone who follows Ligue 1 closely enough understands the unforgivable fault that the public display of a form of proximity would constitute for the two rival presidents. So, many Internet users looked for signs of a deception.
Track down inconsistencies
The first instinct is to count Al-Khelaïfi’s fingers. Are there really five? For a long time, AI stumbled over digital anatomy. Disappointment tinged with worry: there are indeed five of them. Then, we look for other details, inconsistent logos, clothes with an improbable cut or phones with an absurd design. But no, no small visual subtlety betrays the usual improvisations of the algorithm.
Nothing, apart from the incongruity of Pablo Longoria’s empty gaze. A little air of the uncanny valley, this theory of the Japanese roboticist Masahiro Mori which states that the more similar an artificial human is, the more an inconsistency, however slight, seems disturbing to us. Except that the photograph is authentic.
Signed by David Niviere, it was taken during PSG-OM on September 24, and published on the Icon Sport agency website. Other photos taken from another angle, posted online on the Getty images platform, confirm the hug. The author of the latter, Xavier Laine, specifies on X that the shooting took place during the pre-match. Marseille’s honor is safe: Pablo Longoria knew nothing of the loss his team was going to suffer during this “classico”.
An exhausting continuum between true and false
Despite everything, the photo leaves the unpleasant impression of an image that is too shady to be reliable. In question, the general context of the explosion of generative artificial intelligence, and their corollary, the derealization of the photographic medium. This is the materialization of one of the fears of AI observers. Its main danger is not only to deceive the human eye by making the false appear to be true (after all, image retouching already existed in Stalin’s time), but also to make the true appear to be false.
In this, it disrupts our relationship with authentic images. On the one hand, the threat of AI imposes distrust as a default posture. On the other, it reorients our gaze towards a hypercritical attitude, as if, a desperate buoy of a helpless eye, the truth of a cliché no longer rested in its coherence, its grain or its context of publication, but in the meticulous study of points of detail as futile as the number of fingers.
It’s the beginning of endless brain knots. In one of the Getty images photographs taken at the same time, Nasser Al-Khelaïfi apparently only has three fingers. Doubt arises. The other two are actually in the shadows. In the age of AI, there are no longer true or false photographs, but a disturbing, exhausting continuum, where the border between the authentic and the artificial are lost in the gray of algorithmic aberrations and effects. optics, which sometimes produce exactly the same oddities.
This is the first counter-effect of the democratization of artificial intelligence: we will never trust each other again. In this regard, AI also poses a formidable risk of desensitizing the public. What will it be like tomorrow when it is no longer just a simple football photo, but a photo of a natural disaster, of war, of a human tragedy?