AI Art Is Transforming the Creative Industry
If you’ve ever come across an article on QMIND Technology Review, you would have seen corresponding graphics that match the content of the article. Below is a screenshot of some recent articles. Can you guess which ones are made by humans, and which ones are made by AI?
The answer: only the last two are human-made, the rest are generated by AI software such as DALL-E or Stable Diffusion. Did you get them right? Give this to your friends and see how they do. I’m almost certain that not all of us can easily find the AI imposter here.
I’ve been practicing art since I was able to hold a pen. I know the intricacy that comes into creating a piece of art. So when I first heard of AI-generated art, I was pretty skeptical. How can computers possibly understand and replicate the emotions that human art delivers?
This is why I was so astonished when I was able to generate a fairly (or very) decent-looking graphic in DALL-E by simply feeding the software a line of description. On top of this, the AI-generated graphic had all the elements I asked for cohesively integrated.
Is AI-generated art considered stealing?
If you want DALL-E to create an image that mimics the style of Monet, all you have to do is drop the name “Monet” in your prompt. The AI can draw in Monet’s style because it has been trained using a significant amount of Monet’s paintings. Given enough training, the AI will be able to create great pieces of artwork that look as if they were created by the same people who spent years training, perfecting their techniques and finding their niche of styles. Is this fair? Do the AI art generators copy, or steal, other artists’ work?
Yes, but so do those well-respected human artists.
In art, there is a fancy name for copying/stealing — taking inspiration. In most cases, an artist’s work will inevitably be inspired by something, directly or indirectly. This subconscious accumulation of seeing and remembering art from other artists will, at one point, become apparent and trigger inspiration. However, artists rarely cite these names — you never see a bibliography attached to a piece of art. The process of supervised learning makes the artwork the machine is learning from so transparent but isn’t this the machine’s way of “taking inspiration”?
“Good artists copy, great artists steal.” — Pablo Picasso
I do believe, though, that we need more regulations to specify what pieces of art can be used as training data, and ways to eliminate the possibility of AI generating inappropriate art.
Will artists be soon out of jobs?
The current pricing for DALL-E is $15 for 460 images (it generates four images for every natural language prompt). This is way more affordable than hiring a graphic designer. If QTR had to pay graphic designers for all our graphics, we would soon run out of our funding.
Does this mean that AI will soon replace all artists? Maybe not.
Suppose we look back on history and take cameras as an example. The primary school of art during the pre-camera era was realism. People valued fineness in art, take a look at the drawings from the Renaissance period. Therefore, when cameras were first invented and were able to capture objects exactly the way they look, the mainstream thought was that painters would be put out of business. However, this was not the case. Rather than there being fewer and fewer artists creating paintings, what we saw was the emergence of new genres of art, such as surrealism, post-impressionism and modernism, in which abstract art is included. Artworks of these genres can hardly be imitated through photography.
As we can see, cameras are certainly a disruptive technology to the painting industry, but they did not diminish the need for painters. Instead, they created a shift in how people approach art — if capturing objects with perfect accuracy is made very easy because of cameras, what else can we do those cameras cannot? Similarly, AI art might shift our values away from “Does this image have all the colours, compositions and styles that I want?”, to “Is this image meaningful or special in some other way?” Perhaps, in the near future, the increasing presence of AI art will lead to a major movement in the art industry.
We might even be able to see some completely new careers in this industry because of the growing usage of AI art generators. In fact, people often debate whether photographers are really artists since their “art” is created using technology. But professional photographers nowadays are just as admired in the creative industry as traditional painters are, and there’s a need for photographers specialized in certain styles, such as micro/macro-photography, creative photography, etc. AI art generators are analogous to photography in their early state — few people would call AI-generated images “art”.But perhaps as we continue expanding the possibility of generative AI, we will find people working in careers that require more creativity than just giving AI a line of prompt.
Things that Generative AI cannot replace
As powerful as Generative AI now seems, it has some inherited limitations.
First is customizability. Even for experienced AI engineers, the Generative AI system remains like a black box. That is, we only know the input and output. How the system reaches its decision, we don’t know. Therefore, AI art generators might not be our best choice if we want the image to be exactly the way we want it.
Second is the experience. Creating art itself is a fun process that many people wouldn’t want to let go of. The therapeutic effect of filling your paper with nice colours, the sense of accomplishment after finishing that last essential stroke, and the frustration when cleaning up the mess from your paints — you can’t get these experiences from only waiting in front of your computer for the image to come out.
From the creative industry to the whole of society — how should we co-exist with AI?
Rapidly evolving AI technologies are disrupting not only the creative industry but also many other disciplines, such as advertising, finance, healthcare, etc. This alarms many individuals regarding their decision on choosing a promising career that AI won’t replace. However, rather than treating AI as a scary monster, we should instead be proactive in learning what’s going on with these disruptive technologies, try to adapt to it and get ahead of it. The future is hard to predict. If we are always looking for a definite answer to where our world will shift to make an optimal plan for the future, we will eventually be disappointed as there will be no such answers. The only constant in life is change.
“The only constant in life is change.” — Heraclitus