The controversy was that in order for these generative AI art engines to do their text to image creation they have to acquire, sample, analyze, catalog, and mimic millions of artworks without obtaining or even telling living artists who created those artworks that their works are been used in direct violation to their copyrights for pure profit that never will include them in the payout. There is nothing wrong with having a generative AI art engine that acquire proper permission and compensation to artists in exchange for their support and content, aka "blessings."

Of course, the other part of the controversy is taking jobs away from artists now that non-artists don't need to pay hundreds of dollars, if not thousands, to artists to draw what they want. All they need is subscribing to a service and they can crank out artworks they want.

Also, current legal landscape in US and EU explicitly excludes AI-generated contents from copyright protection. Developers and studios are aware of this and starting to put in their employee contract to ban use of AI generated contents in any of their assets for fear of legal "high-jacking" of assets by third-party for commercial purpose. This can be disastrous to a studio that spent tens of millions of dollars to develop a game or a show only to watch bunch bootleggers came along, took their assets, and ran their own production and become even more popular and earn more money than the original developers.

Of course, I know there are artists out there who use these engines to generate references for their own drawings. I do think this is fine since the final product will be the artist's own interpretation and own effort.

The thing is, that's where the funniest part comes out: none of the AI neuronet basically can draw a unique OC, be it furry, anime or just a regular one. So that's a funny trap there: people can be amused and excited about power of ai, still, it won't be able to perfectly draw one particular character, it'll be quite different; plus there'll always be some mistakes, with fingers for example..
Although, from what I know, there's actually new vacancy is going on thanks to ai: you just have to write it nicely to make ai create a thing for you, and the better prompts you give - the better is result. Do you think that's fair? Can and should it be stopped?

That is a fallacy of these engines. Since most users don't know what they actually want and just love everything cranked out of them, they inadvertently taught the engines that those particular styles, layouts, color themes, and compositions are what they are looking for. It will become harder to get these engines to generate unique styles even with extremely specific prompts since the engines are starting to link all of the styles toward one set of "popular" styles by relational selections. This is actually a statistical fallacy. This is what happen when you let unskilled observers dictate the outcomes of events that require experience and specific skillsets. Eventually, those who are tasked with the prompt will have to engineer the hell out of it to get decent results and even then it's an uphill battle.

But at last, the current legal landscape will stymie the development of AI industry and there doesn't seem to be any real movement by world powers to change the status quo. So long as music and motion picture industry groups want to protect their assets, they will influence the laws to prevent blatant thefts and misuses of their IP by AI engines. Universal Music Group has already made their move on Youtube, enforcing copyright strikes against anyone using AI-generated or altered musics with UMG contracted artists or published musics. So far, Youtube is complying and issuing copyright strikes against a few big name AI-generated music channels. So this is reminisce of NFT craze. The technology will stay and eventually repurposed but their initial iteration will die away since they were all built with a house of cards.

Actually, talking about how it might work.. it basically reminds me newbie artists) instead of creating by themselves, the first thing they do is copying someone else's artwork. (and for some reason I forgot what it calls in a more shorter way xD). But what if ai-engineers would actually try to teach AI to draw just like teaching a usual human being? With theory and practice, but not letting it copy what it may "see"? neither be it style or anything that comes along with copyrights. The only thing it would be able to use is references, but creating art as IT actually feels and sees it. That would probably made them create the most powerful AI-artist, not a timeless pictures generator, but AI-artist itself... Wish I could befriend it.. ^^

That would require full understanding of real intelligence, which is a field that cannot be duplicated at this moment since the collective knowledge of human race has not yet reached the point of deciding what is "intelligence" vs "knowledge" vs "experience" and therefore cannot create anything that mimic human psyche fully. Perhaps in the future, once the scientists have finally deciphered the intricacy of psyche and intelligence, we'll finally see a robotic overlord capable of crushing the entire human race within a day before using our body parts for organic raw materials to construct a vastly superior force to go out and conquer the galaxy one alien race at a time.

that's.. a pretty sad future, don't you think? |D