Those who look may feel betrayed…
A photograph generated by AI is the result of learning carried out from human images, whose forms, features and relationships are assimilated and then recombined. It can produce a novel image, but it does not arise from the encounter of a gaze with the real. Is that creating, in the same sense as the person who takes a photograph?
Anyone who knows photography — the kind that says as much about the object or the being captured as about the person who waited, hesitated, observed, adored, hated — can feel betrayed before an image they believed came from a human gaze and that turns out to be generated by AI. It is neither its beauty nor its originality that is in question, but the absence of the human being whose gaze we thought we were meeting.
A text generated by an AI also proceeds from a learning of what exists. The machine assimilates regularities, structures, styles, then produces new combinations. It can write unprecedented sentences, but it does not know the emotion that drives one to search for one word rather than another, to reveal a story or to invent one. Is that creating, when nothing of what is expressed has been lived or felt?
Writing is an act of creation, a revealer of what we are, a collision of truths. The same is true of clumsy writing, the kind that wants to tell us things without mastering the language. An imperfect poem can carry more humanity than a flawless text. Anyone who knows this can feel betrayed on discovering that an AI was the watermark of a poem or a novel they believed born of a human sensibility.
A fiction generated by AI can, in the same way, assemble images, voices, movements and dialogue until it produces the appearance of a film. But the seventh art is also the fruit of an encounter between a director and actors, the result of a deeply human labour in which technique and mechanics serve the human. What becomes of that cinematic encounter when the beings who seem to embody it have never acted?
Must everything that can be replaced be replaced?
History teaches us that we are not condemned to adopt a technology in every circumstance, nor to systematically substitute a novelty for what existed before. We use the plane or the high-speed train for long journeys, but we still walk, cycle or ride horses. We build marine drones while still exploring the depths with a mask and snorkel or with tanks.
Holograms exist, but would attending a projection of our favourite artist in ten places simultaneously give us the same experience as their presence on stage? The motorboat did not erase the sailboat. The first lets us go faster; the second offers us another way of inhabiting time, of dealing with the elements, of observing nature. The lived experience often counts more than the result, because it connects us to ourselves, to others and to what surrounds us.
Why would it be otherwise for art? The existence of a technology does not mean that we must adopt it in every circumstance, at the risk of betraying our fidelity to what gives value to a work, a walk or an exploration. We can welcome the new possibilities without giving up what we wish to keep experiencing for ourselves. This coexistence could even lead us to rediscover what we thought we had for granted. The proliferation of AI-generated images could make us more sensitive to what distinguishes a work born of a human experience. Refusing certain substitutions is not necessarily refusing innovation. It is perhaps, on the contrary, exercising our freedom to choose what we want to preserve.
But why should we refuse to be betrayed by AI?
There are so many moments when we accept illusion. For example, when an actor plays a role, when a director uses special effects, we accept that the real is transformed. When a novelist invents characters, we do not ask them to prove that they existed. So why react more strongly when it comes to AI? Betrayal perhaps begins when this artifice conceals itself and we believe we are looking at an authentic photograph, reading the words of a writer or hearing the voice of a human being, when what we perceive has been generated by a machine.
The actor offers us an illusion with their face uncovered; AI, for its part, can operate masked. But when this illusion is used to deceive us, does the responsibility not belong to the one who chose to substitute artifice for the real?
What, then, should we say to Roger Job’s pupils? Teaching them to flush out the fake will no longer be enough if their eyes no longer allow them to recognise it. Perhaps we will have to teach them to no longer place their trust in appearance alone, but in those who guarantee its origin. For when we can no longer verify for ourselves, our trust also depends on the word of the person who presents the image to us and on the guarantees they provide as to its origin.
This requirement does not concern only the truth of images. It also touches the relationship we have with works and with those who create them.
For behind a photograph, a poem or a film, we sometimes seek someone — singular or ordinary, a stranger or a close one. And it is perhaps that presence that we must learn not to betray.