Consider me an exception to this “we”. I don’t see use of AI tools as plagiarism, or even close to it. The practice can be dishonest if the asset is not attributed properly, but calling it plagiarism is a category error.
That brings me back to my initial question in this thread.
I mean, knowing more about that would be very helpful to judge properly.
We may have to agree to disagree on this point.
If I paint something with the intent of making it similar to Salvador Dali, but without directly copying a particular painting, is it plagiarism? Do I need to attribute it? Assume I’m not claiming it is a Dali work.
Should I be taking this personally or not? I’m hoping not? But I’d be interested to hear what my possible motives could be!
It’s not plagiarism, but it shows a really nasty aspect of AIs. Exchange the name “Dali” with an artist’s name, that’s alive and works as an illustrator for a living. Let’s call him the famous “Phil Riley”. Phil in my example becomes way too expensive, so the buyers decide to make their own Phil Rileys in Midjourney (“Picture of a blah blah blah car in the style of Phil Riley”). Buyers save a lot of money, Phil has to sell the house. So with this style copying you can do imho really bad things.
Dali probably doesn’t care anymore.
At least the judging process (let’s say for ifcomp) is simplified. There are five different cases:
- I’m indifferent about the game. It could be better, it could be worse. = If this is the AI writing, at least it’s not all it’s made out to be. → 4
- I don’t like this game, it’s badly done. The work of an untalented amateur = At least they obviously didn’t use an AI → 5
- I like this game. It’s well written, well paced, intuitive to use and bugfree. = It could be the AI manipulating me in ways I don’t even notice. The singularity is nigh! We have to draw the line somewhere. → 1
- I like this game, etc. etc. The author promises not to have made use of an AI = Who are they trying to fool here? → 1
- The author is upfront about their use of AI. = They are part of the problem, not the solution → 1
His estate still does. Still in copyright, and will be until 2059. But that’s not the point. A Dali-esque painting wouldn’t sell for Dali prices, because it’s not Dali.
Believe me, I’ve tried to use AI to illustrate a story (game). It doesn’t work. You can’t work with an AI to get what you want, even without specifying the further constraint that it be in a particular style. Someone trying to create a quality product using AI images would soon find it’s not feasible.
That’s giving generative AI quite a lot of credit that I’m not sure it has earned yet. (But I know, I know, “just wait a few months to a year, it’s advancing so fast and inevitably it will soon become better than most human work!”)
Putting the quality, ethical, and legal issues aside, I personally won’t be using generative AI for creative projects because it deprives me of why I enjoy them in the first place. Think about making an art piece: you come up with an idea, you can try out different brushes and styles, you keep polishing it up and see how it improves, and there’s a real sense of freedom and creativity. It’s not just about the final piece I make, it’s the fact that drawing and writing genuinely gives me piece of mind. I’m trying to realize an idea I have, and figuring out just how to do it is part of the fun. I can see my progress evolve over time and I’m making something that really comes from myself. That’s the real value in it; not just the final outcome.
Does anyone know if it’s possible to track down sources of AI generated art?
I generated an image that is so perfect for what I want to do, but the background is so composed I can’t imagine it wasn’t made by an actual artist.
I’ve done a google image search, which seems to catch the idea of my image - a silhouetted person backlit by a window - and I think I found the source of the silhouetted person, but the image is changed slightly and all the other “not quite” hits suggest this is a common thing AI art does well, a “person shape”.
heheh. In my case there’s a real sense of abject failure.
Given the way diffusion methods work, that question doesn’t make sense. It’s like listening to a guitar (or other) solo. What you’re listening to may be effectively a product of everything the musician has listened to before, but you can’t say, “well, that note comes from Eric Clapton, those two bars are from Stevie Ray Vaughn.” There’s no direct relationship between the input and the output.
I’d like to echo the sentiment that interactive fiction is a creative medium wherein the final product is understood to be wholly the product of the author’s mind and skills. I’m not necessarily calling AI “cheating”, but I think a lot of the people attracted to IF came for the “hand-crafted” feel, and I don’t see the potential bigger audience for illustrated IF as being worth exchanging that vibe for. It’s quite possible that that attempted trade wouldn’t even succeed.
I’d also like to echo the fact that many AIs count as image theft. We can argue about whether actual artists and writers “steal” by sheer osmosis from consuming other art and writing and understanding what they like about said media so that they may make good creations, but there is at least a manual effort on their part to try and avoid mimicking others and to insert an ineffable part of themselves into their work. One of my favourite games, According to Cain, not only cites specific quotes (Trinity-style), but specifically has a BIBLIOGRAPHY command. I would not object to some sort of ethical use of the technology on Creative Commons/public domain images, but the fact remains that it seems all current consumer-usable AI image generators not only break the laws of copyright, but the idea of authorial consent. I don’t want the IF community being seen as some sort of lawless wasteland where the authors break these rules as long as they get their pretty pictures. We are a community that has traded sensory pleasures of graphics and music for the deeper pleasures of exploring hand-crafted worlds, and we ought to continue to show that mature sense of allowing sensory “deficiencies” in our works in service of a greater good in the results.
Crowther and Woods first started the idea in Adventure that they consented to modification of Adventure and remakes/sequels/derivative works thereof. You might have heard the story of how Crowther hated the dragon puzzle and exasperatedly told his daughter that it was one of Woods’ additions when she proclaimed it to be a stupid puzzle, but you have to appreciate how his munificence led to that puzzle existing in the first place and Adventure gaining a lot more popularity due to Woods completing it. I suppose the point is that I much prefer these types of reuse because the original author will be able to appreciate them instead of feeling cheated or stolen-from.
An aside to anyone who uses Creative Commons/public domain assets… if you know who the original author is and can contact them, it’ll make their day to hear about how you used it.
We are a community that has traded sensory pleasures of graphics and music for the deeper pleasures of exploring hand-crafted worlds, and we ought to continue to show that mature sense of allowing sensory “deficiencies” in our works in service of a greater good in the results. The IF community is one of the friendliest communities that I have seen towards minorities because the nature of the medium itself leads to more-thoughtful people being attracted to it. I am not about to let those community morals fall by the wayside in the service of mere pictures.
As a reminder… the people whose jobs are getting stolen by this are poor and minorities too. It’s not that using AI for a freeware game suddenly makes you sleazy and corporate… but it validates the sleazy corporations who WILL use it in the future to cheat artists. They’ll be able to point to games like this and say that using AI instead of artists is a normal practice. We can’t ever let that be normalized.
At that point one might consider that AIs have passed the Turing test and are really sentient. And at that point you’d have to ask yourself how different this method of producing artwork is from a person looking at and learning from existing art.
Sure, when that point comes. But it’s not here yet, and I’m not sure it’s as imminent as a lot of people, both pro- and anti-AI, think it is. I think people are really getting ahead of themselves with this.
I’m not arguing that it’s imminent. I’m just saying that if you’re arguing AI art is going to escape the limitations it currently has (because it’s basically a stupid dice-rolling machine right now), you have to wonder how close such a master artist AI would be to sentience. I can’t imagine the creativity of real art can come from something not sentient.
This sounds like good ol’ gatekeeping to me.
What I’m hearing is that you don’t want an inclusive community, you want a community of different races and religions etc., but only if they’re intelligent and thoughtful.
Edit: fixed redundant phrasing
Sorry, I think I misread your tone a little there! I completely agree with you on this.
I just realized that @P-Tux7 created his account 2 hours ago. Is it you @severedhand ? Because it’s okay if you want to post after saying you wouldn’t! Nobody would bat an eye honestly.
If it’s not severedhand, then I’m curious if it’s someone else who made a new account to comment on this thread without receiving backlash? (If that’s not the case, P-Tux7, I sincerely apologize)