AI was supposed to make artists' work easier – it's doing the opposite

We're constantly told by software developers that AI won't replace artists; it will make their jobs easier. But some concept artists say the tech is doing neither of those things.

In the video games sector, it seems many studios thankfully still want artists to create concept art, but some say AI is making that work more difficult and less fulfilling. That's because of a misconception about what concept art is actually about (see our piece on what is concept art?).

Generative AI was supposed to make creatives' lives easier by both speeding up technical parts of many workflows and allowing faster iteration. In theory, AI image generators make it much quicker to 'try out' ideas because artists can prompt scenes and concepts before making their own exploratory sketches.

But that's where a problem arises. It can be hard to avoid those early AI-generated images shaping the final vision – particularly if they were generated by the client who's commissioning the work. In some cases, it seems clients may be forgetting the purpose of hiring a concept artist in the first place.

The website This Week in Videogames spoke to a dozen video game concept artists with experience ranging from indie games to AAA studios. All of them said AI image generation had made their work more difficult, even when the AI art was only being used for reference.

“I'm seeing more and more clients generate something approximating their desired outcome and essentially asking me to make 'something like this',” the Scottish art director Paul Scott Canavan told the site.

“It sucks. This practice absolutely invalidates the entire creative process, in my opinion, and makes my job harder and more frustrating. The job of an illustrator or concept artist is to draw from their years of experience to interpret a brief in a creative way.”

The AI imagery can lead to what's known as 'first experience bias'. People tend to become attached to the first vision of something even if it's not the best quality or concept. Better ideas end up being rejected if they don't fit.

The phenomenon has been experienced by movie composers. Placeholder music, often from another movie, is used during initial editing, and clients become attached to it and ask the movie's composer to create something similar, eliminating the composer's ability to make artistic decisions that fit the tone of the movie. The assignment becomes a task of reproduction rather than creation.

"Those images clients show you have an insidious way of worming their way into your head, and I find I have to do a lot more work to sort of flush the system to break away from those inputs," Jack Kirby Crosby told This Week in Videogames.

Some also think Generative AI is limiting clients' ability to imagine how traditional concept art sketches could look in a final game. An anonymous artist said non-artists had become so used to expecting a 'polished product' from AI that it was now “hard for them to imagine what sketches or concepts might look like later down the line.”

“Executive and leadership implicitly demand being shown 'a final product' otherwise they don't understand what they're seeing,” the artist added.

Over on Reddit, many more artists agree.

“Some give a poor concept artist some generated tripe and go 'this but your own', and then demand the real art look more and more like the AI image,” one person writes.

“It's not just people in control not understanding how the artistic process works, they don't understand how artistic collaboration works,” someone else says. “They view ChatGPT as a replacement not just for the things they can't do, but for anyone who might tell them 'no' or offer a conflicting idea.”

One artist from a “dirt cheap Southeast Asian animation studio” complains that some of the work they now receive involves having to “use assets they generated from some slop bot and half the work on the project has just been making the stuff look less obviously AI”.

The risk for the final product is that the concept becomes “this one idea that this one guy had 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 haven't considered how their concept works, or how it looks from different angles.

When a client generates a scene in ChatGPT and asks an artist to develop it, a lot of the concept is often missing, including any notion of what's outside the frame.

“Even if they are genuine about wanting something only in the rough ballpark, you have no idea what the LLM sourced its base material from (except for the near certainty that they stole it), so you can't draw on references as easily or delve into any kind of creative history or process because there simply isn't one,” one artist points out.

“Every piece of LLM 'art' stands alone, the product of a complex process but one without a forwards or backwards chain to anchor it to anything.”

Want to make art the traditional way? See our guides to the best digital art software and the best 3D modelling software.