More and more newsrooms are using AI-generated visual content, but photographers believe this is unfair.

Illustration: Said Selmanović

There is no place for photographs created by artificial intelligence (AI) in real journalism, says Amel Emrić, a renowned photographer from Sarajevo who now works for the international news agency Reuters. Emrić is particularly critical of media outlets that publish AI-generated visual content—he does not read texts with AI photographs and stops following portals that publish them.

“Artificial intelligence cannot create an original, authentic photograph that can be trusted and that has an author behind it”, says Emrić.

According to him, publishing AI photographs alongside texts  suggests that their authors are not using AI only for that purpose.

“It’s a harmful practice of laziness and avoiding real work. Instead of collecting information in the field with a notebook and pen, they find unverified information online and publish it”, he adds.

Using AI illustrations instead of photographs contradicts truthful journalism, agrees Mirza Mahmutović, a professor at the Department of Journalism at the Faculty of Philosophy, University of Tuzla, who teaches radio and television journalism, photojournalism, and communication ethics. In 2024, he published the book Communication Through Photography.

“Images generated with the help of AI may look like photographs, but they lack their essential value: the guarantee of authenticity”, Mahmutović explains.

From reality to fiction

Because of this, Mahmutović emphasizes, photojournalists are considered witnesses of history, providing visual proof of events—something that cannot be undermined even by framing basis, since the images depict real people and events.

“AI-simulated photorealism no longer guarantees the presumption of truth: an image may appear completely credible, but its physical reference in the real world may not exist at all. This erases the line between reality and fiction”, says Mahmutović.

In his view, publishing AI-generated content raises serious epistemological, ethical, and legal questions in journalism.

“AI-generated illustrations, no matter how aesthetically convincing, do not fulfil the fundamental epistemological function of journalistic photography. They are not the result of practices of documenting and witnessing, but of calculation: products of models trained on large corpuses of images whose authors are mostly unknown, with unclear production contexts and copyright issues”, Mahmutović states.

Publishing AI photographs should be left to social media users on platforms such as Facebook or Instagram. They can experiment with it, but they are not journalists, insists photographer Amel Emrić.

“With Reuters I signed a contract that forbids me from altering any photograph I take. That means that I cannot adjust the light, add elements, or intervene in the image in any way. These are the rules of all three major global agencies—authenticity, originality, truthfulness”, adds Emrić.

Publishing AI images in the media as a replacement for photographs is also unacceptable to another prominent Sarajevo photographer, Almin Zrno, who at the same time notes that we live in “rigid capitalism, with technological possibilities that are growing daily and whose limits we do not know.”

“I think this is primarily an economic moment. Why have a photographer, pay him a salary, per diem, or subscribe to a news agency, when you can ask AI to generate whatever you want? But in doing so, the truth is lost. You may get an illustration or a sketch that brings a smile, but it is far from the truth and from what I was taught journalism and photojournalism are about”, says Zrno.

AI photographs in Bosnian media

Some media representatives in Bosnia and Herzegovina, however, hold a completely different view of AI photographs.

In mid-June the portal Dobojski.info published a service announcement from the City Heating Plant about a temporary traffic closure at an intersection in Doboj due to the installation of hot water pipes, on  accompanied by an AI photograph.

Although the portal and its Facebook page later removed the announcement, clarifying that  that it was AI content, it was obvious that it was not an authentic photograph but generated content—one of the depicted workers’ faces looked unnatural.

“It’s clear it’s not an authentic photograph. Services that provide such photos for local portals are expensive”, says the portal’s owner, Slaviša Pavlović. He cites lack of time, lack of staff, and financial constraints as reasons for publishing generated visual content.

“This way, you get content that isn’t truly free, because you still have to pay for it. Free AI systems make nonsensical photos, but for $20 a month you can get them in bulk, while for $200 you can get originals. In that case, it’s better to pay for a proper service”, says Pavlović.

The portal Buka from Banja Luka also publishes AI content, but in a more selective way.

“We mainly use it for illustrations and cartoons, not AI photographs. If they can complement a text or add weight to it, cartoons and illustrations are welcome. Of course, we put effort into crafting descriptions to get the results we need”, Buka responds.

They add that AI systems are “an excellent tool that is constantly advancing and will increasingly be used in the future”, especially on the social media profiles of media outlets.

“If you have a question for the audience or a poll, AI can be used quickly”, they explain.

Pavlović concurs: “This is a trend that will not stop. I think there will be more and more generated content in the media.”

Using AI to create media content due to financial issues or lack of staff is not justification for photographer Amel Emrić.

“Why are they in journalism if they don’t have any money. I’ve been working 35 years as a photographer. For global agencies, I do complete multimedia production—photography, video, text. Believe me, no one uses artificially created animations or illustrations”, says Emrić.

Equipping texts with AI-generated content is not journalism or creativity. It is not journalistic work, but creating fake content,  photographer Almin Zrno underlines.

“Someone did it for you. It’s just typing in key words and getting everything you need. It may be convenient and cheap. It costs almost nothing, but it isn’t journalism”, says Zrno.

“Don’t take me as an enemy of journalists and colleagues. I am an enemy of laziness in journalism. A photographer must go on-site and be in the right place at the right time. Journalists who don’t go to the field publish a heap of uninteresting content. Those are the ones who rewrite texts in the office and rely on AI illustrations”, adds Amel Emrić.

The Future of photography in the age of artificial intelligence

Artificial intelligence has become one of the central topics of our time. It is developing even as this text is being written, and there is no turning back.

AI content will undoubtedly change photojournalism, says Professor Mirza Mahmutović.

“This transformation will not be one-dimensional. Rather than simply replacing photographs captured with an electronic sensor or photographic emulsion with AI-generated images, we will probably see a splitting of photography’s functions in journalism. In certain domains, such as illustrations of abstract concepts, visualization of future scenarios, or educational explainer formats, AI content is already occupying space that photography has traditionally not covered”, explains Mahmutović.

However, this scenario does not imply the disappearance of authentic photography from the media, because that is simply not possible.

“When it comes to reporting from events, covering crises, reporting from war zones, protests, or various local events, authentic photography remains irreplaceable, not because it is technically superior, but because it is rooted in the principle of presence and responsibility”, says Mahmutović.

Artificial intelligence cannot replace humans, their spirit and imagination in creating photographs. AI cannot generate such content, emphasizes Almin Zrno.

“The intuition or spirituality that I have AI does not have in its storage. That is mine alone, just as no one in the world has my fingerprint. What I create, think, and imagine is my unique mark”, Zrno says.

The future of photography in journalism cannot be challenged by technological progress. Newsrooms will always need good journalists, photographers, and cameramen.

“These AI toys are good for commercial photography, but in journalism they are unacceptable, because they are not true”, emphasizes Amel Emrić.

The question, adds Mirza Mahmutović is not whether AI will change journalistic photography, but what normative framework and education will guide this change.

“AI content, if used without clear labelling, can lead to a decline in trust in the media. But if media begin to use it as an additional, not a replacement form, clearly labelled and placed in appropriate genres, it can enrich the visual language of journalism without compromising its core values”, he explains.

Professor Mahmutović adds that this is why many journalist associations call for clear labelling of all visual materials generated with the help of AI.

“For example, the National Press Photographers Association (NPPA) in the USA clearly states that photography is a fundamental means of documenting reality and that AI-generated illustrations cannot replace the work of photojournalists without compromising authenticity and trust. The World Association of News Publishers (WAN-IFRA) demands mandatory labelling of all AI-generated images by AI models, as well as the development of internal newsroom policies to prevent their misuse. Similarly, the Paris Charter on Artificial Intelligence in the Media stipulates that all AI-generated images must be clearly labelled, as does the statement on AI images by the prestigious Magnum agency”, says Mahmutović.

The audience should not be misled by AI content instead of authentic photographs in the media.

“New generations of photographic equipment and applications are striving to use cryptographic signatures and metadata aligned with the open C2PA (Coalition for Content Provenance and Authenticity) standard to provide proof of authenticity. In this sense, the role of photojournalists is evolving: beyond capturing images, they are becoming guardians of the integrity of digital records, ensuring that their work remains authentic throughout the digital production and distribution chain. Of course, the goal of all these measures is not to determine what is true, but to provide the audience with additional information so they can make informed decisions about what to trust”, Mahmutović emphasizes.

Artificial intelligence: Authentic photography is irreplaceable

“Original photography will survive where audience trust is crucial, and that is in serious news. AI can produce an image of sorrow, but it cannot capture the real gaze of a mother who lost her child in war. Only a photographer who was there can do that.”

That is part of ChatGPT’s answer to a question about the future of photography in the media in the age of artificial intelligence. ChatGPT is one of the AI systems that have brought renewed attention to the issue of generated content in the media.

This text was produced with the financial support of the European Union. The responsibility for its content lies solely with Mediacentar Foundation and it does not necessarily reflect the views of the European Union.