We don’t ban Photoshop because fraud exists. We don’t ban literacy because propaganda exists. We don’t ban calculators because people forget arithmetic.
The "style theft" argument confuses influence with copying, something art has done for centuries. Vague references to "studies" about cognitive decay aren’t evidence unless you’re willing to specify what’s being measured and compared.
AI doesn’t replace talent; it exposes the absence of it. Used well, it accelerates learning, iteration, and execution. Used poorly, it produces slop. That’s not a failure of the tool, it’s a reflection of the user. A retard is still a retard, but it scares people like you that your talent was so cheap that it could be reproducible by a program. Could you imagine though, how useful it is being wielded by competence?
My point with suggesting it here is that this dev could use a stepping stone, or a tutor, or somewhere to start with his designs, as it looks like a series of kids drawings. Templates are useful, blindly flinging generated templates into the game without modification is laziness.
Clearly you asked for me to go into tl;dr amounts of depth.
You obviously don't understand how LLMs operate if you think they operate like artists do. There are technical reasons why LLMs can only remix everything thrown into them, using statistical probability to recreate the things they've seen that most correlate to the text in the prompt. Further, there are no means by which LLMs can improve other than by technological innovations in the methods to make them and their databases. They cannot learn, only use what is in their premade database. (This is because they are stateless, which is a computing term that means they cannot truly remember things for later. They cannot themselves retain data or add to their own databases.) Also the sheer waste of water and energy (the kind of data centers LLMs require enmass are very thirsty and energy-intensive beasts) alone is enough reason to be against LLMs.
I do not think using a LLM to help with character designs would help the artist, as LLMs will generally create quite generic designs and it is easy for a newbie artist to copy any flaws the LLM produces (and even the best LLMs don't really understand some things as they are only statistically averaging things, not thinking about things like what composition or what color palette to use). There are good reasons why artists are advised do more than just copying others' art for practice; it can be a valuable learning tool but they're usually better off mixing it in with other methods of practice. I think the artist would be better served by practicing their craft further via more traditional methods, including but not limited to studying art styles they like.
Honestly just making things that require drawing skills like webcomics or games (the very thing this artist is doing!) can be an excellent way to hone those skills as long as the artist makes deliberate efforts to improve over the course of doing it. I've seen artists working on webcomics that have improved drastically over the course of a decade. I've seen someone working on a game on this very site who started out quite unskilled but improved over the course of the game's development. There are good reasons why a common piece of advice is to just keep drawing (or writing or doing other things that require honing a skill) and just not quitting.
Hilariously, you seem to think I'm an artist. I'm a writer not an artist, but I've paid close attention to my artist friends talking about their craft.
Also I enjoy the uniqueness of unskilled artwork far more than I do the kind of averaged out output LLMs tend to put out. Besides, I've gotten off to games on this site that had worse art than this lol.
To get back on topic, this game is quite hot lol. I encourage the maker to just keep working on it!