Exactly. For example, if you then get a better-paying job in Korea because you now speak Korean, or you translate the book you’re writing into Korean and now have a wider market for it, etc. You can see where I’m going with this. What if you write a kick-a$$ text adventure that someone only sees because it had AI-generated art in it and they commission you (commercially) to write something else? How much of your non-commercial use of generative AI resulted in financial gain in the future?
Again, we’re in danger of widening the scope beyond IF, for which there is no conclusion. So to bring this back to Interactive Fiction, I’ll ask a provocative question - where’s the harm in using generative AI for a personal non-commercial hobby project like creating an interactive adventure (assuming all AI-generated content is appropriately attributed)?
It’s funny, I almost forgot this was/is a thing, it’s been so long since I pirated anything. I used a pirated version of Photoshop for 3 years (in college) during which time I made a lot of really bad art and learned a ton of skills from it. I’ve been a paying customer ever since (12 years now), so I think it was a good trade for them in the long run.
I don’t know if drawing an equivalence between Adobe and Indie Artists is entirely apt. Unlike Adobe, those artists may not be artists three years later if they’re struggling to make rent in the meantime.
Like I said earlier, you do you. I won’t criticize you for it as I understand the appeal. It’s like being blessed by a technological superpower to those of us who are graphically challenged. The immense power of this software is… intoxicating. Just don’t mind me if I instead choose to keep embarrassing myself with MS-Paint in the meantime.
For me, it inherently devalues artists’ work in general, regardless of the project’s commerciality. It’s about the fundamental principle of a skilled person being able to make a living with those skills, and not having that wiped out apparently overnight. (I say this as a non-artist who can barely draw a circle, so it’s not my livelihood that’s currently affected by this, just to be clear.)
The “if you’re not going to pay for an artist anyway, it’s okay” attitude doesn’t convince me. In my eyes, you’re still making a public statement that if you want art, human artists are optional. Far better to make your own art/pay someone else for their art/use public domain art/don’t use art at all (delete as applicable).
I made this post knowing that some people would feel this way, a bit risky for me because I tend to be very sensitive to rejection. I don’t want to defend AI-generated art, but I do want to defend my body of work and myself as a developer. I really like to explore different methods, strategies, ideas, etc., especially when they are culturally relevant or part of a bigger conversation. I am very transparent and I like to show people my process so that it feels more accessible to them. I’m not trying to sell anything, I’m trying to connect with people, so it does make me sad to think that some people will shut me out because of this.
So does that extend, for example, to purchasing something hand-made in preference to something that had been mass-produced? Cars, furniture, ornaments, clothes, etc.?
Does that include music that has been synthesised by computers - maybe not by AI, but by automated non-human processes? Or books/text that has been through an automated spelling/grammar checker? I’m not sure where we draw the line between automated tooling that enhances/supports creativity and that which replaces it. “AI” is such a ill-defined term that it doesn’t really mean any single thing. If I take a crappy photo that I use a tool to enhance or adjust, does that count as making a human photographer obsolete? How much ‘editing’ with AI tooling do I have to do before it’s not my creative work any more.
Again, we’re getting away from the topic of IF, but if your stance is that all AI-generated content makes the equivalent human creator obsolete, if I use an AI tool to generate locations in my adventure game, does it make text adventure writers obsolete (and yes, I did as ChatGPT about ‘typical puzzles in fantasy text adventure games’ out of curiosity - presumably this now means my games aren’t creative any more?)
I know I’m being deliberately provocative because I am genuinely interested in this discussion and am keen to understand people’s reasoning for being on either side of the debate. I have yet to make up my mind on it yet, as I can see lots of arguments on both sides, so I really am keen to hear different perspectives.
Just poking in to comment: Isn’t there some rule about “transformative” art? It’s kind of the same issue as sampling: If I take a beat someone else made, reverse it, change the tempo and the pitch and use 7 notes of it perhaps changing the interval of the fourth note, have I stolen something or made something new but inspired by the original?
I think with regard to musical plagiarism, the courts decided a composer can lift a sequence of 7 notes in succession before it becomes a potential legal problem. Composers actually take advantage of this to pay homage or put an “Easter egg” into a song: The musical Wicked (which is based on a novel which itself is theme and variations on the Wizard of Oz books) has a song sung by the character who becomes the Wicked Witch of the West that references “Somewhere Over the Rainbow” using those notes in sequence but phrased so differently (and like mid sentence so it doesn’t stand out) that someone almost would have to plunk it out on a piano to realize it’s there.
Specifically for Wicked...
It’s the “Unlimited” theme that repeats a couple times, starting in “The Wizard and I” - Elphaba sings “Unlimited, my future…[is unlimited]” and those lyrics land on notes with the same intervals (but not timings!) as “somewhere, over the rainbow”.
I have actually done this. My logo for RSPM includes two figures that I found on “free non commercial use” sites which I recolored and cropped and incorporated into a new image so it wouldn’t look the same in a new context.
I’m of course not making money from the game, but my impression is this is what the AI art algorithms are doing by “scraping” art from the internet. It think the issue is the recipient of the art doesn’t know how much it’s changed from the original, and of course the original artist might have originally posted it online with some kind of copyright - even if it’s a CC-BY that would allow this, the recipient isn’t getting the artist data they could properly attribute.
All of this can be directly existential for many people. Some of us know people who are currently losing corporate contracts in lieu of AI art. We see it happening right now. When those people are in our lives, it’s hard to not circle the wagons. Your use of AI art is an outward sign of tacit approval of AI art in general, even if you personally feel differently. Your game also stands as yet another avid example of the effectiveness of AI art to others who may not have as many scruples as you. It’s a beautiful game, graphically and otherwise, sincerely, but that sort of emphasizes the point. The more conditional a boycott is, the more pointless it is. Resistance is probably futile anyway, but is guaranteed to fail if those attempting it start making exceptions.
I hope that makes sense to you. Peeps just trying to survive.
I take issue with this metaphor. When you are making music with synthesis, it is much more akin to constructing a new instrument by hand from wood to create a specific sound that you are looking for. For different kinds of synthesizers, it’s like having different kinds of power tools on hand while you’re making it. Sure, it makes use of computation, in the same way a guitar makes use of strings, but a person is crafting it and playing it, at the end of the day.
Particular pointing out the music synthesis metaphor because it was followed up with this. The kinds of processes involved with automated spellcheck and grammar check are not comparable to soundwave synthesis and synthesizer programming, either. A closer metaphor would be passing a text file through a program you just coded up yourself to do the spellcheck and grammar check.
Soundwave synthesis hardly is a case where human artists are optional, much in the same way that a guitar being played does not mean the human is optional.
Anyway, that’s some distance from the actual argument at hand, but I’m not going to sit by when someone tries to post the tired, beaten, old stereotype of “electronic artists push the ‘make song’ button and it happens”, and try to pass it off as a valid component of a metaphor.
One of my favorite composers who makes CC-BY music actually detailed that he used an AI algorithm to improvise a saxophone solo in one of his jazz pieces. Of course, it required several takes and editing, so it wasn’t just “push the button and let it go” - an artist was required to vet what the AI made up and decide which improvisations worked best in any given part of the song. (It’s also a lot more appropriate with jazz which is often improvisatory by design.) This is a good and proper use of AI.
I am a working artist who depends on selling art to make a living. And I have no problem with AI-- or with anyone-- riffing off my work or even copying it. I mean, knock yourself out. I’ve had people get tattoos from my images, and that was fine. I’ve had people ask me to help them copy something I did, and I helped. If I came across someone passing off my work as their own, I’d probably kick up a little fuss about it, but I wouldn’t get terribly litigious about it. Why? Because that’s exposure, and there is no bad exposure. Anything that puts me out there more is good (with the exception of being used for political purposes I abhor. THAT I would get litigious about).
I believe art is there to be built on. Every artist has influences that are very clear in their work. Every artist steals and borrows. Art is always a transformative process that funnels the work of others through the artist’s eyes and hands. Some efforts are very transformative; others less so. There are certainly boundaries about what is ethical to do, but using AI doesn’t seem very different than making a collage, which I do all the time, and which is perfectly legal and ethical.
It’s probably a little different for me than for a lot of artists, because AI cannot embroider or quilt or cast glass or do any of the things that make my work unique. And it probably will not be able to do these things in my lifetime. So there is absolutely zero chance that it could do anything other than copy my style in a 2D rendering. So I certainly do not speak for all artists. This is just my opinion about use/misuse of my own work.
I’ve used Midjourney, and I found it a surprisingly artistic process. Getting your keywords right to achieve an effect you want is a very creative endeavor. Asking it to make something “in the style of” a particular artist is really interesting, because sometimes it’s clearly just copying, and sometimes it’s really transformative.
If I’m not massively enthusiastic about this particular use of AI, it’s specifically because of the potential for entire professions to be rendered redundant in a matter of a few years, and the economic and social outcomes of that.
I know you’re trying to be provocative, which is fine, but I will say that I think that mass-produced cars, synthesised music and automated spell-checking are all fine, and they don’t change my mind on the preceding point.
Do we have fewer carpenters because IKEA mass-produces furniture? Yes (but we still have carpenters). Do we have fewer sound engineers because we use sophisticated computer algorithms to generate and edit music? Yes (but we still have sound engineers). Am I making a statement that I think carpenters do not add value to society because I buy a wardrobe from IKEA or that sound engineering isn’t a worthy career because I listen to pop music? No. Or am I? What is the acceptable level of obsalescence a new technology or technique can have on a profession? If I believed that buying mass-produced anything, am I acting unethically? And it that’s the cost of ethics, then I can’t afford them, so then the point becomes moot because that’s unrealistic (and further demonises the poor that have no choice).
But to bring this back to IF. If I play a text adventure game that was inspired by ChatGPT or includes art generated by Bing, am I sticking one in the eye of authors/artists that don’t use those tools? And if so, why? Am I depriving someone of a paid commission? Maybe (but unlikely). Am I likely to to deliver a higher-quality adventure game that’s more likely to achieve a larger (non-commercial) audience? Maybe. Will this raise the bar for newcomers to build text adventure games? Maybe. Will this force authors to use AI to enhance their own submissions so that their games get played? Maybe. Will this have a net benefit to the quality of interactive fiction that we enjoy? Maybe. Will it actually just end up flooding the IF sector with low-quality AI-generated content that turns people off playing IF? Maybe.
Keeping this on topic of IF, I’ll repeat my provocative question from earlier - what is the impact of using generative AI in a non-commercial hobby project like interactive fiction (assuming it is appropriately attributed)?
This happens all the time, but we don’t outlaw the internet because it killed the newspaper business.
To those in this thread whose worry is loss of work, you should be advocating for a post-work society rather than a technology-restrictive society.
I see phrasing here and there in this thread that suggest that people don’t understand how these generative AIs work. They do not take an original, or even a set of original artworks, and then change them in some way. Rather, they learn, for instance, what qualities make a picture of an apple a picture of an apple, and then use a rather abstract process similar to solving linear equations to construct a picture out of random noise. If I did this by hand, I’d never be prosecuted, because there’s no restriction on me taking measurements of a picture of an apple, even if the picture is copyrighted, as long as the picture is available for me to download legally. (And if you post a picture to the internet you have granted implicit license to download it.)
So if I can do it by hand, why can’t I apply computing power to do it faster? Legally, it will be interesting to see someone try to stop places like Midjourney.
Ethically, the question is far more murky, I’ll admit. But I don’t think it’ll be stopped by legal challenges.
I personally use free tools and just accept their shortcomings. I use Paint.NET, Inkscape, GIMP, Slate and MyPaint occasionally. I do miss Photoshop and Illustrator though, but I’m not serious about graphic art anymore.
I remember, back in the day, when Photoshop really hit the scene. Every interface had to have bevelled edges and drop shadows. It was abused and everything looked like Fisher Price plastic toys, but eventually people learned to not rely on it as much. I feel AI will be the same way. It shouldn’t be seen as a end, but that of a means.
I think I’ll be one of the skulls being crushed by the Terminators though…
If your goal is to really connect with others, this is exactly how you do it… with honesty and openness.
I wonder if one could pass an art class using Midjourney. The problem with AI art generators is that you can’t really direct them beyond a few words (which they often mix up anyway). You can’t achieve subtleties of mood, you can’t specify fine detail beyond generalities, you can’t get repeated pictures of the same person. There are so many limitations of AI art that just aren’t easily overcome by the diffusion models currently available, that it’s not really viable for commercial enterprises.
