What will become of photography in the era of generative artificial intelligence? What will become of photographers, replaced by text prompts and “fine tuning” AI models that cost much less than a professional? And what will happen to photography exhibitions, cameras, the role of photography as a tool for narrating reality, photojournalism and everything we associate with the world of photography?
Like many other sectors, photography is also looking with concern at the developments of MidJourney, Dall•E, Stable Diffusion and other generative AI capable of creating images so realistic that they seem “real”. In doing so, she is called to ask herself existential doubts as profound as they are unexpected.
Point is the new Italian Tech blog that aims to explore these and many other questions: it is an observatory on the evolution of the world of photography and creativity in the era of generative AI.
Shutterstock conversation and images generated with AI: the copyright belongs to no one, everyone’s earnings from our correspondent Bruno Ruffilli 14 October 2023
The research I want to carry out in this space will start from three (more or less) fixed points:
1) To paraphrase Mark Twain, reports of the death of photography are greatly exaggerated. Every time a photo like the one of the Pope with the white down jacket goes viral, the chorus of cassandras rises up: photography is dead! photographers will no longer exist!
No, that’s not the case: that photo, like all the photos we have seen so far, including those that have managed to sneak into competitions (sometimes winning category prizes), are interesting and surprising images but they are not “photographs”.
Etymologically they are not, because they are not “writing with light” (photography). They are not even conceptually so: they lack what the philosopher Luciano Floridi calls “semantic capital”, all that human influence that generates the stratification of meanings of a photographic image.
AI-generated “digitographies” are something else, something we have yet to fully understand.
To put it with Roland Barthes, the images generated by AI are able to simulate the Studium, the set of references that make the social and cultural interpretation of the photographic image possible, but they lack the Punctum, the fundamental wound that moves the viewer in deeper and more personal way.
2) Artificial intelligence is an integral part of photography and has already radically changed it. Not only that: it made it better, more accessible and more effective. Machine learning and deep learning are the basis of computational photography, the set of algorithmic processes that allows our smartphones to take images of otherwise unattainable quality given the intrinsic physical limitations of the lenses and sensors with which they are equipped.
In the wake of the computational revolution, professional camera manufacturers have taken note. Professional cameras on the market today use AI algorithms to improve key elements of the shooting experience, particularly subject tracking and autofocus. The hardware evolution and the transition from reflex cameras to mirrorless cameras did the rest. Today’s cameras are a concentration of very high technology capable of extraordinary shots, which free the photographer from the weight of technique and enhance his expression and search for meaning.
In short: it has never been so easy to “photograph”, and the credit also goes (in part) to artificial intelligence.
AI has also revolutionized another fundamental phase of the photographic process: image editing and selection. Software like AfterShoot allows you to process and prune thousands of images in a few tens of minutes, tagging out-of-focus images or with subjects with closed eyes, or identifying and grouping similar or duplicate shots. They are tools that save hours of work for the photographer and integrate their workflow, leaving the professional with the final say and the possibility of reviewing the camera’s choices. Ask any wedding photographer what it means to save hours in “culling”, as the process of selecting shots is called in jargon: very few will tell you that you were better off when you were worse off.
3): Generative AI is a new creative tool whose uses and applications (including artistic ones) must be understood and explored with an open mind. Framing generative AI as the absolute enemy of photography or the creative arts is a position as reassuring as it is dangerous, which places us outside of change and does not allow us to understand it. The critical elements that we have observed so far (the problem of copyright of photos and content used for model training, for example) are important and must be analyzed in depth, but they pertain to a broader level than the “Generative AI” problem: here we want to avoid easy classifications to observe and understand the potential of “digitographies” and other visual expressive forms that generative AI can enable.
Within this (large) perimeter there is a lot to explore, observe and understand. That’s what we’re going to try to do with Punctum: keep us focused.
To follow Punctum you can bookmark the main page of the blog or follow me on Instagram (@andreanepori) or X/Twitter (@andreanepori)where I will publish new articles.