The Brooklyn Rail

The problem is not that the current crop of seemingly photographic images generated by algorithms are being mistaken for genuine photographs, even if the credibility of photographs as documents is thereby diminished. The problem is rather that, up until now, we have been making the evidential value of photography an act of faith. Having been firmly in the dark for so long, we find that what AI is doing in this respect is removing the blindfold from our eyes and confronting us instructively with our own naivety.

Algorithmic “photographs” are no longer born of chemistry and light but of data and darkness. And, in fact, they have been with us for at least two decades, albeit confined to a professional or specialized sphere. What is new now is that they are indiscriminately accessible, easy to use and refined in their results. So, we should be grateful that they are proliferating and are within everyone’s reach, because they will remind us of the need to be skeptical, to doubt.

AI-generated images confirm that every image is inevitably an illusion. Let us turn to Elogi de la mentida [In Praise of Lying] (1928) by Josep Torres Tribó, a Catalan libertarian thinker who was murdered in a Nazi extermination camp. In line with the philosophy of suspicion, Torres Tribó advocates lying in an extra-moral sense, as an anti-dogmatic expression of freedom, transcending a mere transcription of the real and manifested in language and artistic creation. Words, like images, do not represent the world but reconstruct it with fictions. We should be careful to clarify the space between extra-moral and immoral lies.

Torres Tribó, a technology enthusiast, predicted that, when machines replace humans in the basic functions of life, we will finally reach the horizon that will foster our full spiritual development. In the uncertain future envisaged by AI, is there still a place for a photography with documentary value or will it become mere illustration? Will the function of photography as historical testimony be reduced to the role played by paintings, such as The Surrender of Breda (c. 1635) by Velázquez and Guernica (1937) by Picasso, for example, or will there be something more?

Both truth and documentary status are determined by use and context rather than procedure. But for die-hard devotees of traditional photographic essentialisms, there is still hope: it is still technically possible to distinguish whether an image comes from a camera or a computer, in the same way that forensic examiners can tell whether evidence has been faked. Ridley Scott’s Blade Runner, inspired by a Philip K. Dick novel, dates from 1982 and presents a plot set in 2019: our time! In this film, “replicants”, robots of indistinguishably human appearance, infiltrate society and can only be identified and neutralized by “blade runner” agents. To distinguish a photograph from an identical algorithmic “photograph,” there is a solution: all you have to do is train blade runners for photography, capable of unmasking these replicant images as well.

The undue attention AI now receives is a matter of media trends, a response to the tsunami of referentless images overrunning us from all sides and the alarm they cause. The development of deepfake technologies has opened the Pandora’s box of the iconosphere, generating millions of hyperrealistic ostensible portraits that are not real. We used to think that seeing is believing. But what happens when what we see does not exist? When what we see is no longer what we see, but a construction, an invention, or worse, a hallucination? The camera also shaped our perception of the world by providing certainty. In a generic way, successive vision technologies have been doing so. Underlying the current bewilderment is the transition from a realism based on optics to a realism based on computing. From Greek realism, Renaissance perspective, scientific illustration from the nineteenth century, and other systems for formatting the gaze, we are suddenly shifting to the sum of all these vision regimes compacted into the computerized automation of representation, which we can apply at will. With a few instructions, a prompt, the computer obediently spits out the images we ask it for.

The revolution has overtaken us and we photographers have to decide which side of history we are on: it comes as a shock that one of these photorealistic, generated images won an important international photography competition. Bewildered journalism spokespersons are wondering whether the time will come when a pseudo-photograph of this kind will win the World Press Photo Contest. But this is the wrong question. What we should be asking is whether a competition like the World Press Photo will continue to make sense in the current scenario. A meteorite fell quite some time ago and we photosaurs acted as though the situation (post-photography) had nothing to do with us. That is, until we came up against this new visual regime brought about by algorithms. Perhaps it is this regime that will finally help us understand that truth is a search rather than a find.