If you look at spookysoulgeek's profile, the posts are mostly calling games AI Slop.
And compared to the average hobby game, "ai slop" will probaby look rather good, quality wise. It's the spammy nature that makes it slop. The slop uploader does not upload a fine tuned small game a month, but often several per day.
Then just limit publishing new projects to one per week.
I dimly remember Itch having a soft cap of 20 projects. Either that was scrapped or they just do not look what type of projects someone publishes, when they lift that soft cap.
What people argue and what they actually do or think, might be two different things. If OP bashes non slop ai games as ai slop, instead of bashing them for using ai, is not relevant for the argument. Actually, bashing or downvoting ai slop is the wrong approach. It gives the actual slop games a sense of relevance. Most games on Itch have no ratings and no comments. Giving the slop games engagement metrics is the last thing any ai hater should want to do.
This thread did not make the argument that all ai assisted games are slop because of ai. I have seen people bring that argument, and I disagree with them.
It made the argument that people are tired of ai. I fully agree with that.
And it tried to villify ai, but no mention of slop or low effort among those reasons.
The idea that AI is not inherently low effort is absurd. If using AI were more effort than not using AI, why would anybody use it?
I mean, I can kind of see a “conceptual artist” spending weeks on an AI stick figure that they could have drawn in less than a minute by hand to make some sort of nonsensical point about generative AI and human effort, but that’s not the kind of project we are seeing on itch.io.
You can claim that the defining characteristic of “slop” is low quality, not low effort, and that high quality can result from low effort. But you can’t deny that using generative AI is low effort.
Don’t get me wrong, I’m much more on the anti-AI side than not, but this isn’t true.
I have nothing against the use of generative AI, for example, for dynamic dialogue, or better parsing in text-oriented games. I’d love to see or work on something like a modern version of Lifeline (2002).
And that’s only in the context of game development. Yes, 99% of AI uses is slop, and yes, 99% of AI engineers and AI companies are interested in enabling slop. I personally don’t have a strong opinion on entirely banning AI in game development as much as in other areas, but in general the above still says nothing of the tech itself.
I said that using generative AI is low-effort, which is a value-neutral statement. It has nothing to do with being for or against AI. The AI bros agree with me. Being low-effort is the selling point of AI. The whole purpose of AI is to save effort. That’s how it’s marketed, and that’s why it’s being used.
If you’re arguing that AI is not low-effort, then that’s an argument against AI, because if using AI doesn’t save effort, then there’s no point in using it. But you’re not backing that statement up with any evidence.
It is your presumption that holds no basis. People spend least effort only within their predefined value system; your fallacy is extending this to a purely materialist stance and closing off any other. I’d love to see your mental gymnastics on why I refuse to use a smartphone, despite it clearly being “lower effort”.
That’s how it’s marketed,
How humans market AI is not value-neutral and those values do not become an innate property of the technology.
and that’s why it’s being used.
That’s awfully reductionist. There is a reason machine learning took over in natural language processing – it is because it turned out to be the better tool for the job, over having billions of hardcoded branches. Plenty of effort went into this development, and I applaud it despite its consequences.
Like I already said, it is true that most people use and develop AI for degenerate reasons, but no, saying “generative AI is low-effort” is not a value-neutral statement, it just reflects your own values on the matter. I gave examples above of where generative AI would not be low-effort. In the case of replacing artists or actors, or delegating everything to an AI agent, it would be low-effort.
Either way, we’ve have both said enough on the matter. I have an actual life to tend to, rather than to argue with nobodies on a dead forum, and I’m sure you do as well.
The idea that AI is not inherently low effort is absurd
No, it is not absurd. Absurd is thinking that claiming absurd things like that will convince anyone.
You peddle a guilt by association concept. Any game that ticks having ai generated content in it is ai slop in your eyes. Is this not so?
you can’t deny that using generative AI is low effort.
Not only can I do that, I can show that this statement is meaningless and not sufficient to argue against ai. You make a simple statement of A always equates to B. That is easily refuteable by bringing a plausible counter example. The argument is also weak, because it calls upon a relative property. Low effort compared to what exactly? You did not establish a baseline of acceptable effort. Anything above is not slop. AI is just not the defining attribute here. Because below that line you will find plenty of human creations. And to top it, you did not even establish that using low effort to achieve a result is a bad thing.
The effort example is this. Consider two games. Made by two devleopers. Let's name them Mario and Luigi, because why not. Mario and Luigi have similar skills and set out to make a similar game. Like competing in a thematic jam or whatever.
Mario downloads a bunch of free human made assets and assembles them together with some gameplay and the theme of the jam.
Luigi also creates a game for the jam with some gameplay fitting the jam, but decides to create the assets with help of ai.
Who put more effort into the game? By any metric I can come up with, it is Luigi. You know, the one that used ai.
Oh, and believe it or not, there is variyng degree of skill and effort you can have and put into using an ai. Just because you use an ai does not make your results publishing-ready.
The slop usually comes out, if a person not skilled in either ai nor the craft uses the ai to mass produce. The archetypical "ideas guy".
Mario downloads a bunch of free human made assets and assembles them together with some gameplay and the theme of the jam.
Luigi also creates a game for the jam with some gameplay fitting the jam, but decides to create the assets with help of ai.
The first problem with that Mario actually did put in more effort than Luigi. Finding free assets that work together and can be used to create a whole game is hard work. Actually free assets are rare. Actually free assets that aren’t AI slop are rarer. Even if Mario is just using an image search, he is still putting in a lot of effort.
Meanwhile, Luigi is basically also using image search, but with a blender attached to the search engine that turns multiple source images into one output image that’s difficult to trace to any single source image. Luigi isn’t restricted to using free input images.
Both are using image search. If Mario uses Google image search, then both are using AI. But Mario is working under the restriction of only using images that already exist and are legally free, while Luigi is free to search a space that includes, but is not limited to, basically all images on the internet, regardless of their copyright and licensing status. Who do you think is going to have an easier time finding images for their game?
But the bigger problem is that you are only looking at the effort or Mario and Luigi, and not at the effort of the artists behind them. When a publisher pays a development studio to make a game, the game isn’t automatically slop just because the publisher didn’t develop the game in-house. The development studio’s effort counts. The individual developers working for the studio count. The third-party contractors hired by the development studio count. And the creators of any stock media the development studio uses, free or paid, also count.
So why don’t the creators of the material that trained the AI count? Because their effort was taken from them without consent and isn’t individually credited. It’s effort, but it’s stolen effort.
So you want to focus on the counterexample. Ok.
The first problem with that Mario actually did put in more effort than Luigi.
Mario did not put any effort into this at all. He just downloaded the assets, while Luigi created the assets. Luigi objectively put more effort into the game by writing the prompts. But selecting the things to use, is an activity you now assert to be equivalent of being "hard work" and "a lot of effort". But if you assert this, you need to concede that Luigi did the same hard work and lot of effort by selecting the outputs of the ai.
Come on. You can use some ethical arguments against ai, or simply taste of the user. But this effort and slop angle is not working.
Actually free assets are rare.
And things like that is why I call out those weak arguments against ai. You need better arguments. Don't overcomplicate it. Users do not like ai, that's simple, true and verifyable. If I verify your claim I get this:
/game-assets/free/tag-no-ai That is 39000+ free assets.
but with a blender attached to the search engine that turns multiple source images into one output image
You conflate different things into this. You try to portrait the internal workings of ai to do a thing they do not. AI usage and how AI work internally is not the thing you describe. You try to construct a strawman to take down. Also, it's a false equvivalence that you construct here.
If Mario uses Google image search, then both are using AI
Image search is not llm. It is not diffusion. It is not generative AI.
basically all images on the internet, regardless of their copyright and licensing status.
You try to load the argument with other arguments, asserting that automatically the ai Luigi used must have images in it's training data that were obtained illegally. We are currently not talking about that. We are talking about effort Luigi and Mario put into a thing and try to decide if we can call Luigi's game "ai slop" based on the effort they put in or because of the tools they used.
So why don’t the creators of the material that trained the AI count? Because their effort was taken from them without consent
the bigger problem is that you are only looking at the effort or Mario and Luigi, and not at the effort of the artists behind them.
It is not about any transcendal or ethical effort at all in this argument. It is about the effort Mario and Luigi put into their hypothetical games.
I did ask, if the following were your viewpoint and you seem to have this viewpoint since you did not object to my wording:Any game that ticks having ai generated content in it is ai slop in your eyes. Is this not so?
You can say you do not like games that have ai in it. But you can not argue, it is because they are automatically slop because of the ai. This is objectively wrong. One does not follow from the other. The only thing that follows from it, is that the game has ai in it. If you dislike ai as a concept and you do not like the game because of that, this is fine.
But do not try to construct weak justificaions for your dislike. Use strong justifications. If you use weak justifications, people can see the flaws in them and decide to disregard your ai dislike as irrational hate.
Mario did not put any effort into this at all. He just downloaded the assets, while Luigi created the assets. Luigi objectively put more effort into the game by writing the prompts. But selecting the things to use, is an activity you now assert to be equivalent of being “hard work” and “a lot of effort”. But if you assert this, you need to concede that Luigi did the same hard work and lot of effort by selecting the outputs of the ai.
Luigi created nothing. But that’s not the point. I’m strictly talking about the amount of effort here.
Write an image prompt, any prompt. Put the same exact prompt into Google image search and a generative AI. Both will give you images in return. Regardless of what happens in the computer, you have done objectively the exact same amount work. The exact same effort.
In either case, you will probably not get something usable on your first try as your first result. You will need to keep hitting that generate button or scroll down the results or refine your prompt. Again, effort. But it’s more effort on Mario’s side, for the following reasons:
This means that Mario needs to put in more effort to get comparable results. Independently of any legal or ethical concerns. Mario could be a cannibal serial killer who voted for Trump and he would still be putting more effort into his game.
For whatever it's worth--
Luigi needs to keep refining prompts, selecting parts of images to change, and dismissing images that aren't what they were after. Trust me, I've done both. Both can be time-consuming, but sometimes AI-generation takes way more work than finding the perfect asset for your project. Also, the former is actually creative effort, not just scrolling and searching through stuff that's already been made.
Yes!
And notice how conveniently the process is shown in a way that only portraits one of them putting effort into it, while disregarding any effort the other put into it. Like the contrived idea of putting an ai prompt as a search term - into image search not less, instead of an asset store that has filters for no-ai and free.
My goal was only to show that an AI user can put more effort into a game than a regular developer. And if it can be constructed either way with individual assumptions, my argument holds true: one cannot infer sloppy development by ai usage alone. Having ai in it, is not enough for a slop or low effort label.
I do not even need to bring creative effort into it, and as you pointed out, there is that aspect. And for that reason, in my opinion, some hobby developers use AI. They can't draw, but at least they can direct the AI to generate something. So it feels like creative input - and it is. It is just not the same input and skill as drawing yourself. Or the same input of searching and chosing the assets - that would be more like curation.
And all three can be done sloppily without even considering skill. Mario just uses the first asset fitting the game jam theme. Luigi copys the theme into the prompt. And Peach, the talented artist, doodles something on the proverbial napkin. Contrast to them doing the same for hours or days, till they get it right.
And notice how conveniently the process is shown in a way that only portraits one of them putting effort into it, while disregarding any effort the other put into it.
I’m not disregarding the effort on Luigi’s side, I’m making the argument that there is less of it, numerically speaking. And I’m backing that up with an explanation. Lots of people use AI as a more powerful search engine. “More powerful” = “less human effort (on the part of the end user)”. And that’s effectively what Luigi is doing: searching the training data for bits and pieces to combine into a new image.
My goal was only to show that an AI user can put more effort into a game than a regular developer.
Well, that’s obviously trivially true. Take any non-AI game. Use AI to add features to it. The new game obviously took more effort than the original game (same effort + extra effort for new features). That’s not an interesting comparison. Bigger games aren’t better than smaller games just because they are bigger, and spending a lot of time and effort creating a big collection of slop doesn’t make any of it non-slop.
What I am comparing is the amount of effort required to get comparable results, because that’s the only context in which “low effort” as a description makes any sense.
And for the record, I think “low effort” is a good thing. I like labor saving devices. I use a compiler that I didn’t write. I use a sprite editor that I didn’t write. I use a washing machine and an automatic dish washer and a vacuum cleaner. I’ll take all of the labor saving devices I can get, so long as I find them sufficiently ethical and they don’t negatively influence the quality of what I’m producing too much.
Weird argument. Both AI generation and human labour can be labour-intensive, or quick/lazy/half-hearted.
If I make a simple game with a few sketches that I quickly threw together, that's not a game I put a lot of effort into, but it's human-made. I once made a game with RPG Maker for the GMTK game jam over just two days. Human-made, but also just a quick project.
On the other hand, I can make a project where I spend hours and hours a day writing prompts for images which I keep refining, and writing scripts with the help of ChatGPT, and so on. Trust me, something as simple as a character portrait can take a loooong time to make using AI if you're committed to it actually turning out well.
