okuha-logo-small-white

Will AI Replace Artists – It’s Looking Interesting

will-ai-replace-artists-feature-image

If you click on a link and make a purchase, I may receive a small commission. As an Amazon affiliate partner, I may earn from qualifying purchases.
Read our disclosure.


Key Takeaways

  • AI art generator creates random art. Artists can create intentional, purposeful, and meaningful art.
  • AI art generators need images to produce art. Without art, it can’t produce art.
  • AI art is not copyrightable, and ownership can’t be claimed. Handmade art is automatically copyrighted.
  • Companies, brands, and studios shouldn’t use AI art as it can be used by anyone in any way they like. This can massively harm the brand.
  • AI art generators are trained with text-image pairs without the consent of artists.

Intentionality

In the context of art, intentionality refers to the deliberate and purposeful decisions made by an artist in the creation of their artwork. It encompasses the artist’s conscious choices regarding the subject matter, style, medium, technique, composition, and overall concept of the artwork.

AI (artificial intelligence) does not have the same level of intentionality in it. While AI art generation is executed through text prompts, the AI decides what to include, how to include, and what isn’t included in the image.

In traditional art (including digital art), intentionality is the artist’s intention or desired meaning behind their work. It reflects their ideas, emotions, messages, or statements they wish to convey to the viewers.

The artist may have a specific narrative, social commentary, personal expression, or aesthetic goal when creating the artwork.

AI, on the other hand, does not have intentionality when creating art. AI art creation is randomized and changes each image generation. This can be seen when the same text prompt is used repeatedly, and the images are always different.

intentionality-ai-example
Even if you use the same text prompt to generate AI images, the AI will always come up with a different image. This demonstrates the randomness of the creation process.

Precision, Mistakes, and Skill

AI-art-mistakes-example
The left image clearly shows a mistake in the hand. The right side image has a mistake in the pupil, and while it’s not so easily spotted, there’s a mistake nevertheless.

AI art generators make “weird” mistakes (as do artists, but not at the same level). Said mistakes can be three hands on a character, missing fingers or extra fingers pointing in weird directions, eyes can be messy, and the face can be disfigured.

While there are many AI art models that have been trained to correct these mistakes, they still often happen.

Artists also do mistakes when they draw or paint, but mistakes are more nuanced, and if you happen to see three hands or seven fingers, they are most often drawn deliberately and intentionally. AI creates these types of mistakes because, for AI, it’s just about algorithmically combining data together.

When artists create an artwork, elements in the artwork come from the artist’s mind. The end result is a precise interpretation of what the artist thought, felt, and wanted to create. This requires skill, knowledge, and thoughtfulness, rarely, it’s the case of creating something randomly.

If AI art generators would use material from their own creation, it would become faulty extremely quickly as it would feed mistakes to itself, which would only create an infinite loop effect to poor results. Artists have knowledge that is gained from years of observing nature, and with that knowledge, they are able to create fundamentally correct-looking art.

Client Work

AI art generation is still a fairly new industry, and it’s expanding rapidly with swift development. However, when it comes to creating art for clients, AI lacks quite a bit.

While you can create AI art for clients, some clients can be picky and want something that is precisely what they have in mind. It could be little details like an earring shaped like a pearl with a black and purple gradient as coloring. AI has a lot of difficulties creating precisely what the text prompt is describing.

The client might require an artist to create a graphic novel starring a female protagonist that has a certain look. AI can’t create this (probably not in a long time) as the face generation is random, always creating closely resembling faces but never the same face.

The mentioned example is only one of many where an artist has the upper hand.

Limitations of AI Art Generators

image-text-pairs-LAION-400M-and-LAION-5B
How image-text pairs look like in the LAION-5B dataset.

AI art generators are trained with datasets that most notably include text-image pairs. These image-text pairs are gathered online (prominently without artists’ consent) and trained to create an AI algorithm able to generate images.

Including ChatGPT, AI models can’t create something that they haven’t had training with. AI art generator can’t create an image of a dog if the art generator hasn’t been trained with images of dogs. Similarly, ChatGPT can’t write about AI art in the same way I can, as it hasn’t been trained with the latest developments in the AI art space.

yoji-shinkawa-ai-art-example
On the far left is an original artwork created by Yoji Shinkawa. The three other images are created by an AI with a text prompt having the words ‘style of Yoji Shinkawa.’ The far right side image is the closest resemblance of the style, created with Getimg’s inkpunk diffusion AI art model.

Artists can create new art styles that are specific to the artist, such as heavy ink by Yoji Shinkawa or clean anime art by Stanley Lau, etc., as they can develop that through years of work and imagination.

AI can’t do this as it does not know what an art style is (it can only be described through artist names that are in the datasets).

Usually, AI artists use text prompts with an artist name, like Stanley Lau, Takashi Okazaki, Greg Rutkowski, etc., just to guide the AI art generator to generate images that look like pieces done by the mentioned artists.

If the art generator does not have text-image pairs from those artists, the art generator cannot create artwork that looks like a certain artist created it.

If artists stop posting their work online, AI wouldn’t have new content to create from and the industry would perish. This applies to ChatGPT and any other AI service dependent on new information.

Copyright And Artists

If there wouldn’t be artists, AI art generators couldn’t produce a single image. AI art generators need images and artwork to create art. ChatGPT can’t write about things it has no prior knowledge of. AI art generators work the same way.

The issue with AI art generators comes from the fact that they’ve been trained with images without consent from the artists. The artists that have published their work online automatically give the right to view (they published it online for everyone to see) the images and artwork but not use the material.

Currently, the AI art generators have been trained with datasets (most notably LAION-400M and LAION-5B) and images that have been crawled (using Common Crawl) from multiple websites such as Artstation, Dreamstime, Instagram, etc.

Luckily sites like Have I Been Trained allow opting out of AI datasets. So if an artist wishes to remove their artwork and images from these huge datasets, that is now possible.

However, this does not limit user-generated models from arising, which won’t ease the artist’s feelings about AI art.

Copyright And Brands

Pure AI art is not copyrightable, and you can not claim ownership of it. If a company uses AI images to promote products or services, anyone can take those images and do whatever they like with them.

This is the very reason why copyright is extremely important in the creative industry. Copyright protects the person or a company so that only the copyright holder can use the images they have created.

Pure AI art does not have this benefit. This is also why clients shouldn’t and most likely won’t use AI art in commercial work, as it can harm their brand if used without consent and wrongfully.

Jobs AI Art Might Replace In The Next Few Years

concept-art-example-image-AI-art
Image with a concept art vibe to it.

Nothing is set in stone. Jobs may be replaced, or they might not be. No one knows until it happens. However, there’s one art industry that might feel a bit of pressure from AI art: concept art.

Concept art is about visualization, ideas, rough concepts, and inspiration for creating a final piece, either film, video game, characters, etc.

AI art is perfect for that type of work as it can produce concepts in seconds, and because concepts can have errors and be messy, the mistakes AI art creates can be acceptable. This depends on whether the company or industry accepts AI generators as part of its workflow.

Regarding graphic designers, there are already vector AI art generators, especially the upcoming Adobe Firefly. The AI art generator still needs input and creative thoughts to generate an output. While anyone can create a logo in the future with Adobe Firefly, coming up with creative thoughts for a logo is still needed.

Incorporating AI to Be Part of The Creative Process

AI can be seen as evil, or it can be seen as a tool that will speed up the creative process. AI art generators can be an asset if used correctly.

AI artists can be an asset in a creative industry by generating ideas, concepts, and visualizations in seconds to be presented in Zoom calls, company meetings, board meetings, etc.

Ai-art-inspiration-example-image
AI art can be inspirational.

As AI can generate images in seconds, it will help companies change creative direction faster. AI speeds up the creation of concepts and gives inspiration for a new work when the artist’s own mind is shooting a blank.

AI images can also be used as reference images. While the images might have mistakes, artists can correct those when creating their own pieces.

As for graphic designers, there are still many people who don’t know how to use graphic design software, so for designers, the logo design process can be accelerated with the rise of vector AI art tools.

Feature image credits.

Search
artist-profile-picture-avatar

Okuha

Digital Artist

I’m a digital artist who is passionate about anime and manga art. My true artist journey pretty much started with CTRL+Z. When I experienced that and the limitless color choices and the number of tools I could use with art software, I was sold. Drawing digital anime art is the thing that makes me happy among eating cheeseburgers in between veggie meals.

More Posts

Thank You!

Thank you for visiting the page! If you want to build your next creative product business, I suggest you check out Kittl!