New generative deep learning models are expanding into all areas of content creation, but they are most disruptive in two of them: text and image generation. For the latter Adobe recently made an important statement: AI is the new digital camera.
The words are those of Scott Belsky, chief strategy officer and executive vice president of design and emerging products of the company, interviewed by the Wall Street Journal: “When artists say to me, ‘I hate generative AI, why do you allow it in all your products?’, I say, ‘Why [altrimenti] people will go to illegal places and you will end up not getting paid at all”.
Then Belsky concluded: “This is a revolution, it’s the new digital camera and we need to embrace it”.
Stock image photographers vs. generative AI
The WSJ article addresses the concern of professional stock photographers, who survived the advent of digital cameras and online sales, and are now threatened by generative AI.
In the piece there are testimonies from some photographers in this sector, who, as happened to Pete Saloutos, they saw earnings from a single photo drop from $2,000 to 2 cents with the arrival of online sales and the growth of supply compared to demand, thanks to the presence of more people who can take photos compared to the 70s, 80s and 90s.
The stock image market has been suffering for some time, if we consider that – in reference to the US alone – the sector saw the Stock Artists Alliance trade association close its doors as early as 2011.
The arrival of generative AI has added to online sales and the increased number of photographers in the market, too the ability to have anyone create imagesnot only to those who don’t have any cameras, but also to companies that usually buy stock photos.
However, there are also photographers who try to keep the bar straight, like Antonio Guillem, who took the famous stock photo of the “distracted boyfriend” on the Internet, and who declared that he still earns good money with the photos taken in his studio, although he doesn’t sell plus the 1600 licenses per day three years ago.
Photographer Tony Northup, owner of a well-known photography channel run together with his wife, said: “I’m an old technician, I’ve been through it a couple of times. So I’ve always been looking forward, looking for the next thing that’s going to destroy all my dreams and the things I’ve built”. A melancholy way to not feel defeated and to ride the new wave rather than be overwhelmed by it.
New rivals: the same stock photo platforms
However, these visions must also deal with another danger: Stock image companies, such as Getty Images, are equipping themselves with tools to generate images. These services go in the direction of being able to count on content that respects author rights (because the models are trained with licensed photos) and which at the same time are more “malleable” than stock photos.
For the moment, this training phase still has “value”. A Memphis photographer told the WSJ that he received a one-time payment from Adobe Stock last year for AI training that amounted to about 10 to 12 percent of his annual revenue from doing business with Adobe. But how long can this value be maintained over time?
What photography is really at risk?
With the launch of Adobe’s new Firefly Image 3 model, which now allows you to generate images from scratch even in Photoshop (they are currently in beta), in recent days we too have wondered about the possibility that tools of this type could harm the market of stock images.
Will AI images replace stock photos? We put Firefly 3 integrated into Photoshop to the test
Go to further information
While the ability to create exactly the subject you want versus finding it in one stock photo among a thousand others is indeed appealing, the results still have so many inaccuracies which at the moment can only have a chance in social media posts seen on small screens and “shrunk” quickly, perhaps to create immediate engagement.
Treating generative AI as the new digital camera is premature and, without a doubt, when this happens it will only be valid for stock photos. As a tool for documenting the facts of life that happen here and now, photography cannot be replaced.
No AI will be able to take the place of professionals who take images of the G7, of war scenarios or of the catwalks of fashion or cinema events. At most, we will be able to think about it when these abilities are reliably granted to photojournalist mini-drones guided by Artificial Intelligence. But it will still take a lot of time. And also very quiet minidrones.
