It has been less than four years since I first saw an image claimed to be created by AI based on text instructions. In a photo generated from the prompt "woman skiing," the figure resembling a human had three legs, and the "giraffe standing next to a tree" was at a level where it was difficult to tell where its head was, how many legs it had, or whether it was an animal or a monster. The gap between language and images, or between machines and human language, was still wide, so it was expected that it would take quite some time for the products of machine civilization to bridge that gap, or that science would acknowledge its limits in catching up with the human mind and give up trying. But that was not the case.
At the Allen Institute for AI (AI2), images generated solely from text captions, such as "A skiing woman on a flat hill" (left) and "A giraffe standing next to a tree" (2020), were created. At the time, AI2 stated that this was closer to a proof-of-concept experiment rather than a machine learning algorithm.
Looking at the AI technology and generative AI images we are witnessing now, hyper-realistic fake photos are already threatening the order of reality. The side of falsehood has gained an absolute ally. It is not unfounded that many scientists worry that within this century, AI could cause a crisis surpassing even nuclear war. Compared to the speed of technological advancement, modern time feels slow. What about art? AI systems like 'Midjourney' instantly create contemplative abstract dystopias that seemed likely to come only after unfortunate artists had painfully experienced devastation and despair and poured their lives into their work. It is somewhat disconcerting that AI even mimics human despair and emotional desolation and expresses them more sharply.
The "Space Opera Theater," which won the grand prize in the digital art category of the Fine Art Competition in Colorado, USA. It was created using Midjourney. (Photo by Jason Allen)
It was expected that after the development of combinational abilities in the realistic domain, AI would move on to the abstract domain, but the reality was different. It is an emotionless state that transcends emotions. The distinction between fact and abstraction does not matter much in such systems. Rather, the mixture of fact and abstraction was the key to breaking limits. Technology has now acquired creativity that surpasses humans. The system nonchalantly produces things believed to be possible only with the human brain, such as the nuances of complex language, metaphor, and symbolism. Machines have become capable of creating what they have not experienced, and humans will only change after experiencing it.
Scenes of Trump being arrested and lost in thought in prison created with Midjourney. Although not many people believe these images are real, they still circulate widely on the internet. (Photo by Elliot Higgins)
Artists have never been free from exposure to various worries and despair. The only fact that can be stated is that artists are facing confusion as a new material for art. Have photographers lost the basis of their profession because photography has lost the trust in its realism, or has the removal of the shackles of proving realism opened up a wider path?
Famous photojournalist Michael Christopher Brown created AI-generated photos of Cuban refugees in 2023. He made a documentary series titled "90 Miles," depicting the journey of Cubans crossing the sea from Havana to Florida. He faced strong criticism for these photos. (?Michael Christopher Brown)
At least, it is true that classical photographic realism can no longer be used as the 'rhetoric' of truth. As with most things, crisis and opportunity have come bearing similar faces.
'Woman Skiing' and 'Giraffe Standing Next to a Tree' created by inputting the same text from 4 years ago using Adobe's generative image program Firefly (? Heo Younghan)
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.

