Software developers always tell us that AI will never replace artists. It will make their job easier. But some concept artists say technology does neither.
In the video game space, thankfully many studios still seem to want artists to create concept art, but some say AI is making that job more difficult and less rewarding. It's because of a misunderstanding of what concept art is actually about (see our article on What is concept art?).
Generative AI was supposed to make creative work easier by speeding up the technical parts of many workflows and allowing for faster iteration. In theory, AI image generators allow artists to “test out” ideas more quickly by presenting scenes and concepts before creating their own exploratory sketches.
But that's where the problem arises. It can be difficult to avoid the initial AI-generated images that shape the final vision. Especially if those images were generated by the client who commissioned the work. In some cases, clients seem to have forgotten the purpose of hiring a concept artist in the first place.
The website This Week in Videogames spoke to 12 video game concept artists with experience ranging from indie games to AAA studios. All said that AI image generation made their work more difficult, even when the AI art was only used as a reference.
Scottish art director Paul Scott Canavan told the site: “We're seeing clients come up close to the outcome they want and essentially ask us to create 'something like this'.”
“It's the worst. In my opinion, this practice completely defeats the entire creative process and makes my job harder and more frustrating. The job of an illustrator or concept artist is to use their years of experience to interpret the brief in a creative way.”
AI images can cause what is called “first experience bias.” People tend to cling to the initial vision of something, even if it's not the best quality or concept. Better ideas can be rejected if they don't fit.
Film composers have also experienced this phenomenon. Often placeholder music from another film is used during the first edit, but the client becomes attached to it and asks the film's composer to create something similar, leaving the composer with no ability to make artistic decisions that match the tone of the film. The challenge becomes one of reproduction rather than creation.
“These images show that clients have an insidious way of getting into your head. I realized that we needed to do a lot more work to flush the system out of those inputs,” Jack Kirby Crosby told This Week in Videogames.
Others believe that generative AI limits a client's ability to imagine what traditional concept art sketches will look like in the final game. One anonymous artist said that non-artists are so used to expecting “sophisticated products” from AI that it's “hard to imagine what sketches and concepts will look like in the future.”
“Management and leadership tacitly demand that we see the 'final product' or we won't understand what we're seeing,” the artist added.
More artists on Reddit agree.
“Some people give bad concept artists some generated tripe and say, 'This is mine,' while demanding that real art look more and more like an AI image,” one person wrote.
“It's not just that the people in charge don't understand how the artistic process works, they don't understand how artistic collaboration works,” says another. “They see ChatGPT as a replacement for not only things they can’t do, but people who say ‘no’ to them or suggest ideas that contradict them.”
One artist at a “very cheap Southeast Asian animation studio” complains that some of the work he's currently taking “requires using assets generated from sloppy bots, and half the work on the project is just making the AI look unremarkable.”
The risk with the final product is that the concept ends up being “this”. one This is the idea one man It happened this morning,” one person warns.
The whole point of concept art is…the concept. But many artists complain that clients who come to them with AI-generated images don't consider how their concept will work or how it will look from different angles.
When a client generates a scene in ChatGPT and asks an artist to develop it, many concepts are often missing, including the idea of what lies outside the frame.
“We don't know, even in broad strokes, where LLM got its base material from (other than the near certainty that it was stolen), so we can't easily consult references or dig into any kind of creative history or process, because it simply doesn't exist,” points out one artist.
“Every part of the LLM's 'art' is independent and the product of a complex process, but there is no forward or backward chain that anchors it to anything. ”
Want to create art the traditional way? Check out our guide to the best digital art software and the best 3D modeling software.
