

actual artists are getting fired from their jobs, because the company replaces them with an AI, that was trained on their original art.
Are you sure that’s happening? Under the previous mode of capitalism, what kind of companies were hiring artists?
As I understand it, that isn’t the actual gripe from the general perspective of the artist. Instead it’s about copyright, a concept I fundamentally disagree with. I don’t think it’s necessary, and that the artist’s capacity for prosperity being tied to copyright is a symptom of a bigger problem than being usurped by software.
I think there is good art and bad art. I think there is good AI art (tbh I can’t think of any examples, I just think in principle AI art has the capacity to be good) and bad AI art. I think the relative ease of access skews people’s exposure towards slop. I use the term slop as a descriptor for AI art that is sloppy or wholly derivative; not to prejudge it.
I think perspectives like yours haven’t compelled me to think they are meaningfully different from that of the Luddites, or those opposed to implementing computers in the workplace, etc. I genuinely sympathise with those groups, but ultimately wouldn’t have us go back.
I don’t think my earlier reply came through. I’ll try rewriting it.
AI can add, remove, change or refine input, either text or image-based, either wholly or partially, which may or may not itself be AI-generated. That feature set certainly allows room for genuine, inspired artistic expression. The way you describe AI art is as though it is all created by asking ChatGPT to draw you something. This isn’t the case, and neglects to consider the litany of AI model types that are fundamentally different to LLM’s. Models which are operated by humans directly interacting with them in a range of ways.
Let’s say you’re a concept artist for a movie. After replacing you with AI, how does the company instruct the model in the concept to be represented? If they’re just asking ChatGPT to come up with something itself, then sure - your description applies. And the output will be shitty concept art, and the movie will shittier than it otherwise would be. People might consume it, but it would be a slippery slope towards failure either because a) people don’t like it, or don’t like it enough for it to reach the critical mass required to spread, or b) someone else does the same uninspired and easy job more cheaply or effectively. If you’re an AI-slop consumer, why watch AI slop movies when you can just watch AI slop Tiktoks?
Good art resonates with people not because humans are easily entertained by pretty flashing lights or whatever an AI can churn out, but because of their relationship to a piece of art which is derived from their human experience. Companies have tried to broaden appeal and lower costs by appealing to the lowest common denominator for centuries, but beyond a certain point it is a failing business model. In my opinion, if some companies want to try, let them find out why there are 1000s of AI-generated movie trailers but no movies.
I think that AI can be used for the concept art in a way that maintains artistic integrity and capacity for artistic expression by having someone skilled in representing visual concepts operate the AI tool. That someone would be for all intents and purposes an artist. In essence the artist position would not be redundant; the way their job is done would have changed.