WILL AI SAVE OR DESTROY PHOTOGRAPHY?

Photography started as a technological advancement. As a document or work of art, it has always evolved with technology. 


An example of this technological evolution is the shift from large-format to 35mm film. This change helped make Henri Cartier-Bresson's work and street photography possible. There would have been no “decisive moment” without the move to 35mm film. 


The shift from film to digital was another technological change that transformed photography. I lived through this transformation. While using the now-iconic rolls, I only had to deliver the slides and the job was done. The client would scan the pictures and do any retouching or editing.

The economics are different now, as film and processing have become expensive. A Velvia film roll is around £25–£40, with development and scanning costing another £15–£30. In total, you are looking at roughly £40–£70 per roll. Ten rolls could therefore cost £400–£700. A hundred rolls would break the bank at £4,000–£7,000. It does not take long for a laptop and a digital camera to become the only affordable options.


The downside is that people now expect a digital file, so the photographer is responsible for acquiring and using equipment and digital tools such as Photoshop or Lightroom, as well as cloud storage and hard drives.


For many clients, the purchasers of photography,  the move to digital was about speed more than cost. In a studio shoot, the camera could be tethered to a laptop, giving the client instant feedback and allowing immediate adjustments. There were Polaroids in the past, but these were often used to check exposure and make sure the camera system was working.


Digital files could also be sent very quickly to a magazine or newspaper. For press and reportage work, particularly when covering an event, speed is everything.

 In 2012, while I was photographing WOMAD, the world music festival that takes place in the last week of July in Wiltshire (UK), I could see photographers from Getty, AP, and Reuters shooting, editing on a MacBook, and sending their work back to the main office. The people at the other end would then send the pictures out to the world. This process is now commonplace across most news outlets, and even more so on social media, where immediacy often takes precedence over longevity. 


For stock photography, did digital technology change things, or was it Getty's expansion, buying up the competition, that lowered prices? Images now come from online image banks. The time of catalogues of photographs is over, and the pictures are downloaded. Regarding AI, several questions arise simultaneously. AI will change photography. But will it destroy it, or change it again?

When photography began, it changed painting and helped spur Impressionist painters. Photographers, while replacing painters for catalogue, fashion, and architectural images, also freed them from the need to provide an exact representation of reality.

Will photography survive AI, or will it drown in an ocean of generative images? Which parts of photography will survive? Street photography that people do for enjoyment will live on. People will still enjoy looking at people on the street, and at the combination of shadows, gestures and odd occurrences.


People also react badly to AI slop, which can damage brands. Part of the problem is the character of AI images: a sameness, or averageness. The AI-generated men and women are attractive but too even, perhaps because they are averages of thousands of good-looking people.


Will there be a premium, or simply an interest, in non-AI images? If you want an image of iron mining in Brazil, it does matter that it's a real image of iron mining in Brazil. Replacing it with an AI-generated image would be the same as using an image of iron mining taken anywhere else in the world. It would not only be void of authenticity, but also damage the credibility of the publication.

The same is true for travel or a personal blog. Do you want an image of an actual church in Prague, or is any vague, made-up church good enough?
 There is also the idea that people will want something created by a human. Will this become a niche luxury, like Leica, Rolex and vinyl?


Instagram and Facebook are already flooded with AI images. As more and more of them become AI-generated, AI will increasingly learn from a world already full of generated images. You end up with an average of an average of an average. 

Yes, AI will give you whatever you ask for. A person walking in the street, wearing blue and listening to music on generic headphones? No problem. But AI cannot give you what you cannot imagine. It won’t give you what you haven't seen yet. 



The other day I was photographing a Samba school parade in Reading. The dancers looked striking in their feathered colourful costumes and charismatic smiles. It was only after I was selecting the pictures that I noticed what I could never have predicted or asked for through a prompt.  I noticed the innocent, happy face of a smiling child with face painting and a colourful beaded costume, holding the hand of a samba dancer. The warm expression they shared gave a sweet impression that complemented the scene rather than contrasting with it. 



This leads us to another question: is art discovered, or pre-imagined? AI cannot find the beauty in nature or the chance occurrence in the world in quite the same way, because, at the moment, AI is not in the world. It only knows the world through what we give it. If we have not yet found something, AI cannot simply go out and discover it.

If excellence or imagination are not required, then AI will be enough. But it will not replace the great photographers of the past nor the great photographers of the future.