Photography and Artificial Intelligence

Are we now facing the end of photography? Will generative images supplant photographs and all their functions? These are just some of the questions posed by the inevitable emergence of AI in all areas of our lives

It is worth remembering that photography, since its emergence around 1827, has undergone a constant evolution in its technical nature. It is enough to recall some of its initial names: salted paper, heliography, daguerreotype, calotype, ambrotype, ferrotype, albumin, cyanotype, platinotype, collodion, gelatin-bromide. These names reflect a history marked by innovation and continuous change in the photographic process. Since then, we have witnessed technological quantum leaps with the introduction of digital backs and the current incorporation of AI. The existence of photography has been marked by significant changes in its production process, which have disrupted its technical and discursive nature uninterruptedly in less than 200 years.

Nowadays, faced with the invasion of AI in every technical and cultural process, some questions concerning the present and future of photography, understood as an instrument of expression, discourse on reality, representation and documentation, are exploding.

It is worth asking: Is this the end of photography (yes, now)? Will generative images supplant photographs and all their functions? Will we be caught up in a banal and empty visual rhetoric regurgitated by generative image programs? Will the generative image achieve the documentary value of photography? Are we facing the end of (human) creative imagination and originality? Will photos be able to be taken without light? Is this not a contradiction?

I remember that sterile debate about whether we had reached the end of photography with the arrival of the digital back, which lasted almost a decade in the 90s. Are we facing a new sterile debate? It will be, as Scott Gray (founder and CEO of the World Photography Awards -WPO-Sony) says, who, when asked specifically about how the competition is dealing with the arrival of AI, answered: “The photo is a photo and AI is AI”. Will it be that simple?

Let us not forget what happened with the image of Boris Eldagsen who won the prize in the Creativity category, in the same WPO-Sony, with the image, considered by the competition as a photograph, entitled Pseudomnesia: The Electrician, and described by Bing as “an intriguing exploration of the intersection between human creativity and artificial intelligence technology in the world of photographic art” (Bing, 2024).
Boris Eldagsen. Pseudomnesia: The Electrician, 2023.

Boris Eldagsen's image reveals the potential of AI to generate images that look like photographs, and opens the horizon, Pandora's (black) box, of what is possible in figurative or abstract representation, documentary, informative or artistic function, non-fiction and fiction of the generative image.

This reconfigures (as they have been reconfigured throughout history) the values of interpretation that we have about images. Likewise, it imposes a redefinition of the potential values of photography as a representation of facts and realities, the need for new social pacts of reading and a new role of the image in the construction of human memory, and what this entails in terms of the configuration of identities.

Museums and galleries are progressively receiving more "photographic" works made with artificial intelligence. It is worth remembering Rogelio Séptimo's work, Exhumar la memor.IA, winner of the Acquisition Prize of the XX Photography Biennial in 2023. In this work, the author recreates some images from a photographic archive that never existed, combining analog photography and generative software, forming an individual and collective memory.
In this confusing moment, where AI begins to dominate the scene, photography experiences an accelerated process of hybridization that presents features of continuity and rupture with respect to its previous processes and discourses, both analog and digital.

Some features of continuity of photography in generative images are basically related to the discursive and narrative, with the form that the visual representation takes. AI is, at this level, a machine for remediating forms that already exist. It is also a simulation of the forms that light creates when recorded on a photographic medium. But AI does not work with light, it works with data (machine learning and deep leaning) and is random, predictive, stochastic, and statistical.

Some of the groundbreaking features are precisely that AI does not work with light or space, it is not referential. Working through text or voice commands (prompts), it still requires a human operator, but it synthesizes images from massive databases. It is true that a very high percentage of what an AI system does depends on human decisions. But it is also true that these machines are endowed with an indeterminacy index that gives them a growing margin for autonomous learning.

What lies ahead? Without a doubt, we will have to face the changes that AI will put before us in terms of the values we attribute to photography. The only possible path will be to begin a process of literacy about AI and its impact on the culture of visual representation. We will have to relearn how to consume and produce photographs, which will now be even closer to the simple and complex status of images.
Source