Photography and AI: A Fertile Tension

Photography rediscovers an essential function in the age of generalized doubt

Introduction

Now seems the perfect moment to reflect on a striking paradox: never has it been easier to verify the truthfulness of information, and yet never have we doubted so much what we see.

Today, powerful tools are available to almost anyone. Reverse image search, metadata checkers, and AI-enhanced search engines allow us to trace sources and investigate claims with relative ease. But something deeper has shifted. Until recently, people often trusted images by default. They believed in their truthfulness not because they had checked them, but because they assumed they were unaltered. That trust was sometimes misplaced, yet it shaped a culture where the image still held a privileged link to reality.

Now that manipulation is everywhere and widely acknowledged, that instinctive trust is gone. The capacity to verify exists, but the emotional bond with the image has changed. In this new climate, truth is available, but belief has withdrawn. The image is no longer the bearer of proof. It has become ambiguous, open-ended, and increasingly unstable.

In this context, AI reshapes the stakes for both art and photography. While some lament the so-called end of the truthful image, what we are witnessing is not a disappearance but a redefinition of the link between creation and truth. The most common critiques of AI echo outdated arguments once used against photography in the nineteenth century, failing to acknowledge that art has always evolved alongside its tools.

So rather than asking what AI can or cannot do, the real question is: what does photography become in a world where artificial images are everywhere?

The anti-AI critique: a debate already settled in the 19th century

Since the emergence of generative AI, a recurring discourse has attempted to exclude it from the realm of art by posing a supposedly fundamental question: Can AI create art? But this question is based on an intentionally flawed reasoning.

Anti-AI partisans always offer the same first argument: AI cannot create art. That is like saying a camera cannot create art. Of course it cannot. But an artist can use it to make art, because it is the artist that makes the art, not the tool.

Do these people not live in the same post-Duchamp world I live in? I still find it remarkable how people celebrate art from the unconscious, outsider art, or appropriation art, yet become deeply unsettled by AI.

Artists use AI to create art. That is all. And they have been doing so for decades.

So the real question is not whether AI can make art, but whether artists use AI to make art. The answer is unequivocally, definitely, yes.

Before it went mainstream, artists had already been working with AI since at least the sixties. Think of the early generative experiments.

More recently, artists like Ekman Mansimov have sold AI-driven essays to major American institutions. At Paris Photo 2023, this lineage was visibly present. And in 2025, a sale at Christie’s featured several creators I know from X, who presented remarkable work born from AI processes.

Artists do not wait for theoretical validation to create. They experiment, repurpose, and integrate new tools into their practice. Art history abounds with similar moments.

When photography first appeared, it was met with intense criticism, as documented by Paul Edwards in I Hate Photographers! An Anthology (PUF, 2006). Many claimed it could not be considered art because it mechanically captured reality, without the visible hand or subjectivity of the artist. What they failed to understand was that over time, photography would be claimed by artists, and thus transformed into an artistic medium. Today, that is self-evident. A camera has no soul, yet it can produce soul-filled art when used by a human. The same now applies to AI.

History repeats itself every time a technical innovation enters the field of creation. In the early days of photography, people feared that the proliferation of images would devalue art. Some even predicted that painting would disappear, that mechanical reproduction would destroy the aura of the artwork. But photography did not annihilate painting. It transformed it.

Impressionism, and most of modern painting, were born from that confrontation. The new medium shaped new mental images of the world that had never been widely accessible before. Masterpieces gained renewed value, while everyday image-making gradually shifted to the new tool.

Similar arguments resurfaced with the invention of metal paint tubes, accused of destroying academic painting by allowing artists to work outdoors. The same happened with synthesizers, feared by traditional musicians, and with samplers, suspected of killing originality by making copying too easy.

Each time, these fears proved unfounded. Art does not disappear when new technologies arrive. It adapts, evolves, and absorbs new tools. To reject AI on the basis that it is too automated or that it does not truly create is to cling to a view that misses a fundamental truth. Art is not made by tools. It is made by the way artists use them.

Moreover, art history shows that every major transformation responds to a deeper need: the urge to represent the world differently, when existing forms no longer suffice.

As Alkan Avcıoğlu notes, each artistic rupture comes from the need to express what existing forms can no longer capture.

Realism emerged in reaction to Romanticism. It was soon surpassed by Expressionism, once depicting outer reality no longer sufficed to convey inner life. Each era develops new visual tools to shape its concerns. Surrealism emerged to explore the newly discovered unconscious. Postwar abstraction arose to represent what could not be represented: the trauma of war.

So the question is not whether AI can create art. The real question is how it redefines our relationship with images, with reality itself, and with the way we used to relate to it through photography.

AI challenges photography to reclaim its role as a witness

Far from eclipsing photography, AI restores its crucial importance. In a world where anything can be artificially generated, what gains value is what can still be proven real, authentic, and rooted in the visible world. The rise of synthetic imagery does not signal the end of photography. On the contrary, it reaffirms its essential function as a witness.

This shift is especially visible on social media. Instagram once celebrated the polished, hyper-curated aesthetic of the glamour age. Today, TikTok reflects a collective hunger for spontaneity, imperfection, and immediacy. Younger generations, having grown up in the chaos of visual overload, often possess an intuitive sense for what is staged and what is real. Paradoxically, many older viewers are now the ones who mistake obvious fakes for truth. Poorly generated AI content continues to circulate widely, largely because many older viewers still take it at face value.

In this new landscape, the documentary power of photography returns with new urgency. It is no longer enough to be visually compelling. An image must carry the weight of evidence. Photojournalists and documentary photographers now bear an even greater responsibility: to produce images that can withstand scrutiny in a post-truth world.

This renewed pressure on truth may also lead to new tools. Cameras capable of generating cryptographic signatures with every shot already exist. These devices do more than answer a technical need. They respond to a cultural demand: the need to trust what we see. Verification mechanisms such as embedded metadata, blockchain certification, and digital authentication become more than protective measures. They become part of the meaning of the image itself, a philosophical gesture toward the possibility of visual truth.

This evolution reframes our understanding of photographic authenticity. In the past, photography was valued for its capacity to capture reality. Now, it is increasingly valued for its ability to prove it. That subtle shift may define photography’s role in the years ahead, as perhaps the last visual medium capable of making a credible claim to reality.

The same applies to artists. Some will find renewed meaning in producing unaltered files, building entire bodies of work around the idea of purity, of traceability, of physical presence. Others will blur the line on purpose. But both paths gain new relevance, precisely because the artificial is everywhere.

My Work Is Now a Duality in Tension

Between 2018 and 2020, I often felt unable to articulate my photographic work. Not because I had stopped creating, but because I could not find the words to accompany it. Writing has always come second for me, and during that time, the world felt too opaque, too overwhelming to translate into language.

The arrival of generative AI was, for me, an unexpected gift. It allowed me to give form to the images I had only been able to imagine or describe in words. Today, I experience a fertile tension between strictly real and entirely imaginary images. There is a balance at play, a dynamic flow that nourishes my creativity.

I feel more focused and more excited than ever to return to what is specific to photography: being of the world, in the world, bearing witness, even if the image is filtered, interpreted, or simply framed. At the same time, I enjoy building visual fantasies through AI, exploring the latent space of the imaginary.

I have developed a way of prompting that makes my AI images uniquely mine. No reverse description tool could reconstruct them. What defines them is not the descriptive input, but the specific language I use, which is deliberately non-descriptive.

My approach is a mix of guidance and misdirection. I try to create conditions where the system can surprise me. I replicate, in a way, the feeling of walking with a camera, knowing roughly where I’m going, but always alert to what might appear. The key is to jump when something unexpected emerges that could help me build something new.

Someone once explained why he preferred the obedient nature of ChatGPT-4’s image generator over the unpredictable behavior of Midjourney. He said: I want a computer to create what I want it to create, not what it wants. I see it differently. I come from photography, in the traditional French sense. You move through the world. You respond to what catches your eye. Life is surprising, chaotic, unplanned. The art of the photographer is to frame it, to make sense of it, even fleetingly. That is serendipity.

This is also what I seek in AI. When it allows me to mimic the apparent randomness of nature and human life, when it throws something unexpected into the frame, then I feel I can begin to shape meaning. Life flows. You cannot go against the current. You follow it, with your eyes open.

Who is "I"? What is "want"? What is "create"?

If you are an artist, these questions are probably central to your work. And as for the machine: it wants nothing.

After twenty years of practice, my work remains scattered, across hard drives, discontinued platforms, obsolete formats. I now feel the need to bring it together, to give it a tangible and accessible form, both online and in print.

The dialogue between the real and the imagined has transformed my visual storytelling. Where I once felt torn between the limitations of photography and the abstraction of writing, I now move fluidly between the two. My photographs capture moments lived. My AI images give form to mental landscapes and portraits that could never be photographed.

This duality has added new layers of meaning to my work. When I photograph a street scene, the image holds documentary value, but also enters a dialogue with imagination. Conversely, when I generate an image through AI, I rely on my photographic instincts, my eye for light, for composition, for the decisive moment. My experience with film stock, grain, lenses, the frame, and how it feels in unpredictable situations all continue to shape my choices.

Far from diluting my photographic identity, this integration has clarified it. It has brought me back to what photography does best. It holds a unique link to the world. It captures what will never happen again. And it embraces chance, not as a flaw, but as part of its truth.

Conclusion

When I began working with AI, I did so with complete naivety. I had no idea that so many of my fellow photographers would reject it so strongly. Looking back, I am deeply grateful for having followed that naive instinct. For me, it was a form of honesty, the same honesty that has always guided my photographic work.

I never had a plan. I never followed a grand structure. What has always been difficult for me is showing images that feel so deeply true that their intimacy makes me hesitate. But I try to stay faithful to that instinct.

Now that I have reached, as Dante put it, the middle of the journey of our life, or as someone else might say, Mid-Journey, I feel for the first time able to look back and forward at once. I now have the perspective to give meaning to both my past work and what I want to build from it. The two paths are not separate. They run in parallel.