How AI is Reshaping the History of Photography

Like most, I was carried away by the epochal change to pictures that took place circa 2022, once it became possible to generate photorealistic images with the help of artificial intelligence. I read articles about DALL·E and Midjourney, becoming aware of the technology in much the same way a comfortable painter might have learned about photography in the early 1840s, acquiring a kind of knowledge that felt peripheral and required no action.

Not too much time has passed, even if the technology behind synthetic imagery has improved significantly. The scholar and photography critic Fred Ritchin, who began writing about changes in media in the 1980s, has just published an essential primer for mass visual literacy in the age of artificial intelligence: The Synthetic Eye: Photography Transformed in the Age of AI.

The book’s second chapter, titled “Playing with AI,” ends with a historical coda that made me chuckle: “Many of these early synthetic images are like the daguerreotypes produced soon after the invention of photography, accused by Baudelaire and others of being ‘art’s mortal enemy.’ The critics were right, as many 19th-century painters undoubtedly agreed, but also quite wrong.”

The Synthetic Eye is interspersed with synthetic images “created by the author via text prompts,” as noted at the end of the book, made “in collaboration with either OpenAI’s DALL•E or Stability AI’s DreamStudio between 2022-24.” Indeed, of the 88 illustrations, only one, at the beginning of the first chapter—aptly titled “Exiting the Photographic Universe”—was taken with a camera. This is an impressive indulgence. “With both trepidation and enthusiasm,” Ritchin writes, “after several decades spent editing, curating, and writing about photographs, I began to experiment with generative artificial intelligence systems that bypassed the camera, hopeful that the images produced in response to my text prompts might be freer and more innovative, without some of the restrictions I had experienced.”

The restrictions Ritchin describes relate mostly to photography’s troubling inability to illustrate what is outside the frame. Though the technology powering photography has changed significantly—lighter weight cameras, DSLRs, Photoshop, sharper lenses, smartphones with front-facing cameras—its images are still indexical, traces of what is or has been there. Bypassing the camera and its constraints became possible only because the technology of making photographs has produced a surfeit, with an estimated 5 billion photos produced daily, mostly on smartphones. That is, these cameraless, synthetic images are progenies of those camera-born ones.

No criticism about images today can evade the question of improbable scale. In fact, it has become somewhat uninspiring to bemoan the image glut—a practice in criticism that began, at the very latest, in the early 1900s. Now, it is essential to speak of the reality created by an avalanche of images. This is the gist of Ritchin’s propositions in his two final chapters, where he advocates for “a responsible use of artificial intelligence” that does not “simulate the photographic,” but helps to “explore the questions provoked by these images… pathways of inquiry that AI supplies to amplify and interrogate the photographer’s work.”

One such question concerns modern warfare. With battles increasingly fought with drones, is the war photographer’s role resigned to a post-event witnessing of the trauma suffered? If so, will we come to accept a camera-made photograph of a devastated landscape side-by-side with a synthetic image of a built-up city as illustration of the cost to rebuild? And since photojournalists are now routinely targeted, can we imagine a war entirely reported through synthetic images?

As the photographic diminishes in scale, this era of “meta-photography,” as Ritchin terms it, means that an image can serve “as a portal,” a tasking attempt to “investigate what lies behind” the image. These investigations reach the sharpest degree of their moral dilemmas when it comes to the suffering, pain, or imagining of others. There are few qualms over photorealistic images of a cup on a table, but great kerfuffles follow a fake image of the Pope or of Rafah. It is not an overstatement to argue that the human need for reality, even if muddled by a virtual world filled with post-truth technologies, remains as intact as when the camera obscura was invented.

Ritchin’s closing suggestion is, rather than supplant the indexical, “We can then use artificial intelligence to investigate what is outside of photography’s ken, and also to make sense of the trillions of images that have been made while, within constraints, helping to conceptualize what they depict.”

Photographs, he is ultimately saying, will not disappear, even if they become scarcer. We must prepare for a transformative “age of AI” in which boundaries blur between synthetic and camera-born images, in which we are obliged, as everyday critics and engagés of visual culture, to make sharper distinctions between them. Failing which we are damned by our illiteracy.