(20th July 2025, 1:36 PM)Weltall Wrote: I think there's a lot of reactionary reaction to AI based on a dramatically narrow viewpoint of what AI even is. Furthermore, it displays staggering closemindedness and categorical rejection, which is unbecoming of imaginative and progressive people.
If you use a LLM to write for you or to generate an image, and you call it your own, you are fraudulent.
If you use a LLM to help you write your own material, it can be a tremendous creativity multiplier. You can brainstorm with it. You can use it to edit. You can use it to collect your notes and categorize all the elements so that it is all easily accessible. If you use an LLM to help you visualize an idea, or to show you examples or what a specific object looks like in a dozen different styles, or you want it to generate color palettes, all to help you create an image based on an idea you generated from your brain, that makes it a wonderful tool.
Furthermore, to be perfectly frank, I don't think much of the argument that it "steals" from art, because I think that point of view is almost certainly an astroturfing campaign stoked by the likes of Disney and other media conglomerates, who actually do stand to suffer from AI generated knockoffs of the intellectual property they own. And, tbh, I don't feel bad for them at all. All of Disney's keystone IP at this point was either acquired from the original artists, or created by original artists who have mostly been long dead, and you and I all know they'll be using these tools themselves without a moment's hesitation if it will save them bux. Small time artists aren't being stolen from or hurt in any meaningful way, unless their "art" consists of sort of consumer-grade slop designed to be printed on plastic wrappers that end up in a landfull. AI can definitely do that.
I am not saying there aren't legitimate concerns, but some people have developed a legitimate phobia about the subject, and it will not be to their benefit.
As to the former, using AI to generate an image or written piece and claiming it as one's own creation, I completely agree. Unfortunately, that is what a lot of students in writing-intensive classes are now doing. I suppose it's not terribly different from when they used to plagiarize from Wikipedia or from online repositories of student essays, but AI-generated writing is harder to catch. I have strategies, but even they aren't foolproof, and I'm not entirely sure how much I trust AI detectors (especially when they flag Grammarly as AI usage).
I also agree about the many valid uses for AI, such as brainstorming, editing (I don't have a problem with Grammarly, personally, and it is getting more accurate, though still not a complete replacement for proofreading), and organizing of ideas. I use ChatGPT for these purposes sometimes, whether it's organizing ideas for a new class syllabus I'm designing or generating lists of fantasy-sounding alchemical ingredients or tertiary character names for my creative writing.
A lot of independent and prospective up-and-coming writers and artists (and hopefuls) are concerned about theft of their work being used for "training" the AI programs, and I generally try to avoid sharing specific details from my stories with AI programs like ChatGPT for that reason. That being said, as an unpublished writer, I realize I'm much too obscure to influence the algorithm in any meaningful way, so if anything, I'm practicing an abundance of caution, perhaps unnecessarily. Still, though anecdotal (and no one I know personally), I have heard of instances of artists having their work that they posted on deviantART copied (or seemingly copied?) by AI programs, so I don't know that they're only learning from Disney and other major media conglomerates.
Again, I don't think AI is necessarily evil. I don't like the ways it's forced on us (for instance, being the first result that appears in a Google search often at the expense of accuracy as it tries to synthesize data from, say, some guy's Reddit post), and I don't like the way some people use it (some of my students using it to cheat on their essays being a particular frustration for my colleagues and me; to a lesser extent, I am disappointed by independently published writers who use AI-generated images for cover art instead of hiring a human artist, but that can sometimes be more the result of economic barriers). That said, AI does have its uses. I am concerned about the environmental impact of overusing AI--whether or not its energy consumption has reached (or ever will reach) crisis levels is something that remains to be seen. Nevertheless, I do tinker around with ChatGPT quite a bit, so I'm not one to virtue signal. Ultimately, the grievances I've expressed have less to do with AI itself and more to do with how people are using it.