Photojournalism in the Age of AI: The Importance of the Photographer's Presence
Rotem Rozental
Generative AI introduced a new feature into photojournalism’s complex operational system: the production of speculative, fictional images of events that could have been. In March 2023, Eliot Higgins, the founder of the open-source investigative collective Bellingcat, posted images on X depicting then-former President Donald Trump violently arrested, being dragged away by police officers in one, pushed in another, or running away (Trump! Running!). The caption read: Making pictures of Trump getting arrested while waiting for Trump’s arrest. The images quickly went viral, triggering widespread concerns in legacy media outlets about the exacerbation of misinformation and the flooding of our communication platforms with the new reality of deepfakes. The fact that those images were created by a respected investigative journalist added to the anxiety expressed in what amounted to thousands of words that highlighted how technological advancements also pushed the limits of our relationship with truth, its capturing and distribution across media platforms.
By mid-March of that year, Trump’s legal predicaments intensified. He himself anticipated being arrested on possible charges brought by the Manhattan district attorney related to the Stormy Daniels hush-money case. Higgins’s wishful thinking generated images that portrayed a possible reality, and his authorial voice as a truth-seeking journalist, combined with the hallucinations of the machine, produced a fictional hybrid that went beyond deepfakes made by hate-mongering trolls and keyboard warriors. They seemed like a documentation of a near future rather than an exaggeration designed to go viral. Higgins’s images felt new, and their reception complicated further questions around the dynamics of viewership and participation at the dawn of the age of AI. And, by the way, as to be expected, the images were riddled with easily detectable mistakes (Trump might just have three legs and an officer’s baton strapped to his waist in one).
Higgins’s inadvertent statement about journalism’s relationship with the truth and the present in the age of AI was added to a much longer discussion about the production of misinformation and its transition from the depths of the darkest corners of social media to the heart of mainstream discourse.
In 2021, the same year OpenAI introduced DALL-E — the first popular text-to-image generative model — photojournalist Jonas Bendiksen released The Book of Veles, which included an introductory text and portraits that read and looked human-made but were in fact generated by GPT-2. The fake portraits of the residents were created with 3D modelling tools, combined with real shots of the city that Bendiksen captured in 2019-2020.
Bendiksen turned his attention to Veles, a town in North Macedonia, as it gained notoriety for becoming a central location for the production and distribution of misinformation and disinformation online, designed to tilt the US election in favor of Donald Trump in 2016. With limited financial possibilities, Veles’s younger residents found a new industry of remote labor to support them.
Bendiksen’s project was a response to the rise of fake news in our political systems, expressing a worry about the operational methods of channels of production and distribution of information, and our inability as viewers to separate truth from manipulative fiction that is fed to us as truth. Bendiksen, however, did not tell anyone (other than his publisher) that this was a work of machine-generated semi-fiction. The book was released as a traditional documentary project but was in fact a well-designed hoax that revealed our lack of preparedness to second-guess what reliable sources provide us. The Book of Veles both tapped into a pivotal moment in our changing relationships with images and added a conceptually challenging chapter.
Here too, then, the insider voice is key. Bendiksen was a celebrated photographer and a member of the prestigious Magnum Photos agency. We expect him, in other words, to accurately depict real-life events.
Bendiksen himself, so it seems, was hoping someone would catch him after he released snippets to social media in advance of the publication. But that did not happen. It also did not happen when the project was introduced as a video screening during the prestigious photojournalism festival Visa Pour L’Image in September of that year, a screening which Bendiksen himself avoided. Realizing no one seemed to note his actions, Bendiksen left the festival and soon thereafter turned to fake accounts he created on social media to let the world know something was wrong in the project. It was only then that Twitter could no longer ignore it. The jig was up, as well as the curtain that unveiled the digital and virtual truths behind the scenes.
In a statement posted on Facebook, the festival’s founder, Jean François Leroy, apologized to his community, indicating that he and his team were “flabbergasted” and “dumbfounded,” but, he added, “soon started thinking about the lessons learned from the experience.” And here, again, the question of trust—not in the image itself, but in the person who promises its status as fact —arises. Bendiksen, Leroy recognized, was a “regular” at the festival. In other words, they had no reason to doubt him. Fake news, so it seems, was now finding its way to the heart of the photojournalistic discourse in new ways.
These anecdotes articulate how, in this moment, photojournalism reveals something significant about our expectations as viewers: that even when we know that Trump does not have three legs, we desperately want to believe the image, while it desperately shows us repeatedly that we should have a more nuanced understanding of what it says about the world. And beyond the image, we want to believe its creator. And that trust was deeply destabilized since the introduction of large language models (LLMs).
Now, the more exposed we become to hyperrealist images that are not anchored in an event that actually happened, the more we doubt everything we see. Our relationship with images—and in particular with images that aspire to capture the world—is eroded. And that erosion happens during those few seconds we spend asking ourselves whether what we see is fake. Our critical visual literacy is now a social survival skill.
Image manipulation has been integral to photographic practice since its early days; a history that should be acknowledged and considered in the context of our current relationship with images. In that sense, fake images and pictures that stretch the limits of our vision are nothing new.
As early as 1840, the French photographer and inventor Hippolyte Bayard superimposed himself in a staged image he titled Self-Portrait as a Drowned Man, depicting Bayard as a suicide victim to protest a lack of recognition in his work. In the 1860s, William H. Mumler discovered by mistake that double exposure could generate images that seemed to capture ghosts—an error that brought him financial success until a man depicted as a “spirit” was identified as being very much alive. During the American Civil War, field photographers working for Matthew Brady were staging bodies to increase the dramatic effect of their compositions.
The digital era also took those histories into account: since the introduction of photo editing software in the 1990s, attempts to tell the story of photography as a technology of manipulation and material agency rather than a mute tool designed to capture truth gained in popularity, culminating with an exhibition titled Faking It: Manipulated Photography Before Photoshop, shown at the Metropolitan Museum in NYC in 2012-2013. But since then, digital media became the main source of news and smartphones became ubiquitous globally, the fakes have become deepfakes, and our relationships with images became more confusing than ever.
It’s obvious that the authority and position of photojournalists today is more complicated than it would have been a decade ago. Photojournalists are now required to use the same digital tools that destabilized our relationship with the truth in order to produce a record of a truth; maintain an authority while everyone assumes their smartphones provide them with the same authority; compete with AI-generated imagery, and still maintain a trustworthy voice, somehow.
We’re used to thinking about photojournalists as witnesses, recorders of reality. But what if we reposition that assumption, and consider the photographer as a participant, whose presence both guarantees the event has happened and redefines the event itself? What if we consider the photographer as the generator instead of the chronicler? Does the photographer’s presence matter more now than the image itself?
Christopher Anderson’s White House shots for Vanity Fair landed at the heart of photojournalism’s existential crisis, right in between the viewers’ lack of trust in visual information and photographers’ struggle to recalibrate their profession in the age of AI. Anderson captured Vice President JD Vance and the President’s top aides and advisers, including Susie Wiles, Stephen Miller, Marco Rubio, and Karoline Leavitt, up close and personal. No retouching. Just the bare truth of their faces: blemishes, wrinkles, recent filler shots above the lips, eyes wide open, raw, in color and black and white. They seemed saturated and intense, confident yet extremely vulnerable.
In later interviews, Anderson shared that Stephen Miller was extremely anxious about getting his portrait taken. As they were finishing the session, Anderson told The Washington Post, “he [Miller, RR] comes up to me to shake my hand and say goodbye. And he says to me, ‘You know, you have a lot of power in the discretion you use to be kind to people.’ And I looked at him and I said, ‘You know, you do, too.’” In response, the Internet was quick to suggest that Anderson realized his power and captured a portrait of Miller that seemed to reference portraits of Joseph Goebbels, one of the most important ideologists and propagandists within the Nazi regime and its Third Reich. Whether we accept the comparison or not, Anderson seemed to have understood the magnitude of the access he was given and how critical it was to maintain the authenticity of his photographic style.
In 2017, Anderson photographed Trump himself in a similar style for the cover of The New York Times Magazine: a close-up in black and white, deepening into the smallest details of his skin, shown behind Trump’s own quotes about other politicians, perceived enemies and peers. “My hope was to make an image that feels intimate and unlike the caricatures we normally see of him, good or bad,” he told the New York Times Magazine at the time, “one that somehow communicated the experience I had of meeting him.”
Throughout the 2010s, Anderson’s camera tried to pierce through the veneer of a glossy distance with multiple politicians, including Barack Obama, Bernie Sanders, and Ted Cruz. The proximity renders their eyes, apprehensions, or facial gestures inescapable. We see them as they might not expect to see themselves.
Beyond a direct statement about the politics or decisions made by Trump’s advisers themselves, Anderson’s images from the White House in 2025 struck a chord precisely because of the nuanced documentary perspective they presented and the poignant reminder that the photojournalist holds power, since he does not mute or turn off his subjectivity. The photographer articulates a point of view of political figures, which is then subjected to further editorial decisions. Anderson’s portraits seemed to have told viewers something genuine about the people in the photographs—about the gap between how they see themselves and how they are captured by the camera; how they want to be seen and how, the images seem to tell us, they will inevitably be viewed, not as strong movers and shakers, but as politicians that seek the approving eye of legacy media.
“It’s shocking to me that the world expects reality to be removed from a picture. My intention is not mockery or cheap shots. I’d like to think I’m a stone-faced but critical observer,” he wrote on Instagram following the immediate response and intense conversation the images generated in late December. Anderson’s documentary style succeeded in crafting a defiant statement about photojournalism, precisely when it seemed to have been under threat. We see and sense his choices, his account, his presence on-site with his subjects in each physical imperfection. And when we see the contours of a conversation between a photographer, his camera and their subjects, the photographer’s agency feels restored, vital and pressing.
We didn’t need Anderson in the White House to give us a contested objective truth. We needed him as a chronicler, to tell us something significant that would not be visible to us in any other way. His portraits matter not because they are not retouched, but precisely because they provide us with a subjective point of view and involve choices that remind us the photographer can be a witness—that something had happened that needed to be captured, and that the truth of the encounter, of the photographic dialogue, still trumps our doubting eyes.
Author
Rotem Rozental, Ph.D., is the Executive Director and Chief Curator of the Los Angeles Center of Photography. Between 2016-2022, she served as Chief Curator at American Jewish University, where she was also Assistant Dean of the Whizin Center for Continuing Education and Senior Director of Arts and Creative Programming. Her book, Pre-State Photographic Archives and the Zionist Movement was published in 2023 (Routledge) received the Jordan Schnitzer First Book Award from the Association for Jewish Studies.
Rotem is a lecturer in the Critical Studies Department at the USC Roski School of Art and Design. She mentors artists worldwide and contributes regularly to magazines, journals and exhibition catalogues. Her writings were published in Artforum, Photographies, Jewish Currents, Artillery, Tablet and Forward, among other outlets. Recently, the essay “Call the Midlife,” taken from her current book project, was published in Afterimage, and her short story, “A Scar, Visible Only to Me,” was featured in the anthology Scars by Beyond Words Magazine.
Recent curatorial projects include the group exhibitions Reservoir: Loneliness, Photography and Well-Being (2026, catalogue), Glowing Earth: A Century of Photography, Immigration and Resilience in Little Tokyo, Expand and Contract: AI and Alternative Processes (2025) and If Memory Serves: Photography, Recollections and Vision (Brand Library and Art Center, 2023–2024). Recent solo exhibitions included Ali LeRoi: Everybody Hates Exhibitions (2026), and Matthew Finley: Lost and Found (2025).
 | A guest post by
|