Revolve / Maison Meta

Editorial Photography & AI

© Revolve / Maison Meta

Will AI mark the demise of photography? What challenges do photographers face, and can the medium endure? Join us as we delve into these critical questions.

─── by Josh Bright, November 20, 2023
  • In April 2023, photographer Boris Eldagsen sent shockwaves through the photography world when his image, titled ‘The Electrician’, claimed the top spot in the Creative Open category at the World Photography Organization’s Sony World Photography Awards. However, Eldagsen declined the award, revealing that his creation was not a photograph but, in fact, entirely generated using AI.

    Ai generated B&W image of two women by Boris Eldagsen
    'The Electrician' © Boris Eldagsen

    Eldagsen claimed he used the image as an experiment to challenge the competition and ignite a discourse about the future of photography. Some questioned his motives, perceiving it as a cynical publicity stunt, but whatever your viewpoint, this event unquestionably reverberated throughout the photography industry and prompted important questions about the role of AI.

    While AI-generated imagery has been in existence since the 1960s, with pioneers exploring the potential of computers in art, it is only in the past decade that we have witnessed the emergence of truly realistic imagery. This evolution can be largely attributed to the rise of ‘deep learning’, a subset of machine learning that utilizes multi-layered neural networks to gather insights from vast datasets.

    AI generated image of a corgi dog on a beach
    Image of a corgi created using DALL-E © Courtesy DALL-E

    You might wonder why, given the availability of high-quality AI imagery for over a decade, it is only recently that it has garnered such attention. The primary reason for this can be traced to the development of OpenAI’s deep learning algorithm, CLIP (Contrastive Language-Image Pretraining) in 2020. CLIP seamlessly combines natural language processing and computer vision, enabling it to effectively comprehend and analyze the connections between words and images, thus laying the foundation for the creation of AI art using text-based prompts.

    Further acceleration occurred last year with the advent of Latent Diffusion Models, a form of AI technology capable of generating creative and visually appealing art. OpenAI’s DALL-E, in particular, emerged with the unique ability to transform text descriptions into images, turning words into visual representations.

    Image of the pope in a big white coat. Ai generated
    The viral AI-generated 'Pope in jacket' © Pablo Xavier

    This surge in popularity has brought forth a proliferation of ‘realistic’ AI-generated imagery, such as the recent viral depiction of the pope adorned in an incongruous winter jacket.

    However, this rapid rise has also raised ethical concerns. AI-generated ‘photos’ can be employed to spread conspiracy theories, propaganda, and other nefarious purposes. And, while photo editing and manipulation have been part of photography since its inception, a distinction exists between traditional editing and the creation of entirely AI-generated images, prompted by individuals.

    AI generated image of Donald Trump being arrested
    Ai generated image of Trump's arrest © Eliot Higgins

    Undoubtedly, the lines have blurred with the advent of AI tools, such as those available in Photoshop which allow users to select, remove, and replace elements within their photographs. The technology for producing photo-realistic images is, for the most part, still constrained, there are often subtle details that expose the image’s inauthenticity, particularly when human hands are involved, which often have an unusual appearance.

    Ai generated photo of a model with flames behind her. From an article on "photography & AI"
    © Cultured Mag / David King Reuben
    The world's first billboard with AI generated image
    The 'world's first' artificial intelligence (AI)-generated billboard campaign © Revolve / Maison Meta

    Nonetheless, AI is an ever-evolving field and presents a challenge to ‘real’ photographers with commercial brands increasingly likely to use AI-generated images instead of photographs. AI creative studio Maison Meta has created a host of advertisements and campaigns for fashion brands, including the ‘world’s first’ artificial intelligence (AI)-generated billboard campaign for retailer Revolve, whilst well-known fashion and art bimonthly publication, Cultured Magazine, recently published a 30-page editorial with Louis Vuitton exclusively featuring AI-generated images.

    An AI-generated image of a female portester being arrested by police in Colombia
    An AI-generated image used to depict protests and police brutality in Colombia. Amnesty International claimed they used AI-images instead of photographs as a measure to safeguard the identity and security of the protesters © Amnesty International

    Amnesty International, recently faced criticism for employing AI-generated images of Colombian protesters instead of using authentic photographs, a troubling incident involving the world’s leading human rights organization, whose ethical standards one would expect to be higher. We’ve even witnessed some well-known ‘pure’ photography platforms endorsing such imagery. While this has sparked inevitable backlash, it appears that they anticipated this and are willing to lose some of their core followers in pursuit of a broader audience captivated by this new technology.

    At times, one cannot help but feel that AI-generated images are attempting to deceive us. While some AI creations venture into the realms of science fiction or surrealism, clearly delineating themselves from reality, others attempt to mimic more typical photographs, with increasing success (particularly to less discerning viewers).

    AI-generated Image of Mahatma Gandhi
    AI-generated Image of Mahatma Gandhi which recently went viral © Jyo John Mulloor

    Take, for instance, portrait artist Jos Avery, who amassed a substantial Instagram following with his monochrome portraits, which he presented as genuine photographs. He captioned the images with the names of the subjects and anecdotes about them and even went as far as to provide details of the camera he purportedly used. However, eventually, apparently overcome with guilt, he decided to disclose that all these portraits were AI-generated through Midjourney and then further edited in Photoshop.

    He may have revealed his secret, but questions still linger about the integrity of his approach. For instance, consider the below image supposedly paying tribute to Elouise Cobell.

    Jos Avery's AI-generated portrait made 'in homage' to Native American activist Elouise Cobell. © Jos Avery
    color portrait of Elouise Cobell
    Photographic portrait of the real Elouise Cobell © Fire in the belly productions

    The model-like portrayal in no way resembles the Native American Activist, as pointed out by one commenter who remarked: So this is supposedly an “homage” to an Indigenous female born in 1945-ish and your ‘artistic impulse’ was to make her look like a Kardashian-esque mixed race model!? Where is the original photo, if any exist, of the actual human being? And if you’re so inspired by her, why not just post a photo of her as she actually was?

    While this may be a single example, it epitomizes the approach of numerous AI ‘artists’, who make images that supposedly mirror ‘reality’ but in practice, are artificial and idealistic, driven by the quest for accumulating the highest number of likes.

    The rapid proliferation of AI could also be of particular concern to photographers who have captured images possessing an otherworldly, surreal quality or a level of timing and framing so perfect that it almost defies reality. Take, for instance, the image below, captured by Brian Goldfarb. While it may appear otherworldly, it is, in fact, a real photograph, taken on the Bonneville Salt Flats near Salt Lake City, USA.

    color photo of a man and tv at Bonneville Salt Flats, Utah, USA by Brian Goldfarb
    “The Insomniac”. Bonneville Salt Flats, Utah, USA © Brian Goldfarb

    Goldfarb came up with the idea of shooting a surreal image of a friend watching TV amidst the salt flats while on a road trip in the area. With the help of a close friend, he went to great lengths, securing a TV and borrowing a generator and extension cord (later removed in Photoshop) and in freezing temperatures and high winds, captured this remarkable image as the sun was setting: an example of what photography can achieve through imagination, meticulous staging and substantial effort. In the past, such images would have been celebrated for their mastery. However, today, there is a risk that they may be dismissed as the creation of a well-prompted machine rather than the result of a skilled and creative practitioner.

    Here at The Independent Photographer, we’ve contemplated the role of AI imagery extensively in recent months. We recently issued a statement, reflecting our viewpoint, and received overwhelmingly positive feedback from our community of photographers and photography enthusiasts.

    As we study the captivating work of the masters and are taken back in time, imagining the moments they pressed the shutter as all elements converged for but a fleeting instant, or when we examine the diversity of incredible submissions we receive to our awards every month which transport us to every corner of the globe, we are truly in awe of what this medium can achieve.

    Color travel photography of pilgrims walking in Varanasi, India by Eduardo Ortiz
    "Ghats and souls". Varanasi, India. © Eduardo Ortiz

    While AI-generated imagery may have its place, it can never truly replicate the profound essence of photography. The genuine talent and artistry of those leading the way in our field remain unparalleled, reminding us of photography’s unique ability to immortalise real-life moments that, without the camera, would be lost forever.

    However, this topic is undeniably ever-evolving, and the question we must ask ourselves is not whether AI-generated images meet the criteria of art, but rather, whether we should afford them the same value as those crafted by photographers. Consequently, we must decide whether we are comfortable with the potential scenario where images generated by prompts and machines supplant a medium that demands immense skill, creativity, effort, dedication, and, most importantly – as so eloquently put by one of the medium’s greatest – humanity.


    “There is one thing the photograph must contain, the humanity of the moment.”

    Robert Frank

    All images © their respective owners