AI Art in Games - Discussion Thread

[tebbno]

I haven't really seen anything yet that comes across as anything other than a quick cash-grab or compensating for a lack of genuine artistic talent/effort.

Happy to be enlightened though.

 

[Ratataka]

If the developer is still working on it, it could be good

 

[incognito2009]

The crux is,
Most devs won´t put much work in AI art. You have to put much photoshop hours in good AI atm. If you do it. It looks amazing.
So it is like every game. It depends on the work of the dev.
Lazy dev horrible game. Hard worker dev and creative. Good game.

That represents pretty well how I fell about it, AI art needs to be polished at photoshop to look good, the problem is that most who use ai art are either unwilling to spend that time or simply don't know how to. Which ends giving this uncanny valley to most Ai games.

 

[zezaozezinho012]

Lots of games made with AI, what do you think?
I personally don't like it, I think the characters are very similar

 

[ErzuzB]

I think the characters are very similar

Unlike DAZ, HS or Koikatsu.

 

[thatoneuser]

Not a fan at all of AI art in general. It's boring and usually not better then an actual artist. Same with music made by AI (fucking horrid), or "art" made by AI (also fucking horrid). It's lame, lazy and you won't get the same quality anyway. I avoid AI games like the plague. People who get excited about it (atleast how I see it) are the same people who go crazy for "tiktok/instagram artworks" where a guy hung up some buckets of paint to make patterns, it's not fucking art and it will never be better. Visuals aren't everything in a AVN but I would be lying it it didn't play a big part. But obviously not everyone can actually draw very well, or in a specific art style so therefor unfortenately for now - we will be seeing more AI coming up. Same with those guys copypasting Koikatsu models / animations (or DAZ / anything else that they just yoink), doesn't mean ofc that EVERY Koikatsu game is trash but it does mean that most of them are.

 

[Crpapjda]

I can get around if you tell me that a game is helped by IA, but a game 100% made with IA, no thanks. The art is weird, the stories are bad, and it doents improve any other aspects of the games. Instead of helping the devs to focus on other parts of the games, they just use it to make a fast print game with minimal (to non) effort.

 

[Smrdgy]

They feel like either the author couldn't be bothered, or just wants to do a quick cash grab, or in rare cases, is incompetent with a 3D software. With that being said though, give it a few years to program and train the models properly and it just might get to the point where it's an excellent tool for pure writers to bring their stories to life. For now though, "ai cg" is in my tags blacklist.

 

[AlBetGam]

Lots of games made with AI, what do you think?
I personally don't like it, I think the characters are very similar

Like any other tool, a whole lot of devs will put in the minimum work and get a crappy and generic-looking product out of it. Because the minimum is a lot lower here than it is for programs like DAZ, that "whole lot" will be bigger and the quality floor will be lower.

But there will also be (and already are) some devs who put in real hard work and actually become skilled, and turn out quality games with genuinely appealing AI renders. I'm always happy to play those games.

 

[Karamchad]

AI feels souless, generic and boring.
You can see in the details of the art, the shadowing, the fingers etc.

 

[MattShizzle]

Unlike DAZ, HS or Koikatsu.

Except Koikatsu looks way worse than if it was pictures of actual shit.

 

[bingbear]

AI today can help you with outlines and story ideas. It can do generic art and photogenic women. It cannot put together a coherent story and match the story to the artwork. On some sites it is burying human creators with meaningless noise and reducing the usefulness of these sites. If left unchecked, we will get 100s of new games every hour, all variations of the same game, without soul. Finding a human written game would be looking for a needle in a haystack. And if it was good, then you would find 1000 poor AI duplicates next week.

 

[trc450]

Aversion toward AI is pretty normal since the ease creates excess amount games compared to other mediums type. I'm optimistic, tools are getting better and better, lazy games six fingers all guchi.

Login to civitai.com and browse if you don't like what you see then AI just is not for you. (Recommend models / choose pony and checkpoints only)

 

[perikolui]

AI is advancing by leaps and bounds, in a short time humans will no longer be necessary

 

[Merriadoc]

Yo,
I understand the general apprehension regarding AI-generated Lewd Games, and perhaps art in general, for the variety of reasons that have already been explored at length - be they the lack of quality, the lack of 'human touch', or AI Games becoming the equivalent of Email-spam in its unoriginality. However, it seems inevitable that AI will become, at the very least, another tool for game developers in terms of writing, coding, and indeed art-work. As AI progresses I'm certain it will be eventually indistinguishable from a "human-made" game and after that, simply better. However, we're certainly not there yet, given the state of many of these games...

So, I'm kind of curious what excites people about this new powerful tool for game developers? Personally I'm looking forward to the first games, lewd or not, where NPCs are essentially your ChatGPT (or any other LLM), allowing you full freedom to interact with this character that is confined by their own prompts, giving them a backstory, interests and goals. After that, games where the world itself changes depending on your choices - choices unforseen by the game developer. That would mean that a game developer is no longer the story-writer, but perhaps merely the equivalent of a dungeon-master of a DnD game.

Anyway, TLDR; What are REALISTIC ways AI might be incorporated into any near-future game that actually excites you?

 

[Somnambula]

Bringing characters to life with AI sounds cool.
But then there's the difficulty of branching the story and having assets in the game.
So I think this is mostly realizable only in 3D games with a large open world.

 

[Merriadoc]

Bringing characters to life with AI sounds cool.
But then there's the difficulty of branching the story and having assets in the game.
So I think this is mostly realizable only in 3D games with a large open world.

Yeah, I suppose it boils down to NPC:s being brought to life with AI before the world as a whole is. Sort of a constructed playground but with AI-infused playthings.

 

[thegroan]

I must preface saying I am an AI-skeptic, for me even the term is misleading.
I agree in that it would be nice to have, as you mentioned, an AI as a branching blank-filler, intermediate scene creator, NPC "soul" and world dungeon master. This would happen during playthrough based on input and within the story's constraints. The main drawbacks I see in this approach is the need to be always online and dependent on another company's LLM, which for sure would censor topics, the increased energy cost, and that quality control and testing would be a bitch (it would be very difficult to gatekeep the AI to follow a coherent story)
But more realistically, I see AI as a workload-reducer, but only developer-side: Extend some dialogues, mash some ideas in a coherent manner, translate the game, make intermediate scenes, de-age or age characters as needed without the need to model them again and so on.

 

[Merriadoc]

I must preface saying I am an AI-skeptic, for me even the term is misleading.
I agree in that it would be nice to have, as you mentioned, an AI as a branching blank-filler, intermediate scene creator, NPC "soul" and world dungeon master. This would happen during playthrough based on input and within the story's constraints. The main drawbacks I see in this approach is the need to be always online and dependent on another company's LLM, which for sure would censor topics, the increased energy cost, and that quality control and testing would be a bitch (it would be very difficult to gatekeep the AI to follow a coherent story)
But more realistically, I see AI as a workload-reducer, but only developer-side: Extend some dialogues, mash some ideas in a coherent manner, translate the game, make intermediate scenes, de-age or age characters as needed without the need to model them again and so on.

I'd be inclined to agree with basically everything you said. Any skepticism or qualifiers as to how a game, certainly with today's capabilities, implements AI is fraught with problems. Not least in terms of the censorship that you mentioned. It all comes down to the current technology, as I'm certain LLMs will become cheaper, more cost-efficient and honestly ubiquitous at some point - but the point is that we're not there yet. I suppose the sketicism resides in these half-measures of AI-implementation until the technology has advanced sufficiently.

 

[dereguedeixe]

I like that it can help the development of a game, but when we talk about the art or text of a game, then its a no. I really doubt that we wont be abke to distinguish it since even with all its upgrade you can still easily see how it is ai.


I dont have a problem if the artist use it to draw a city background in those sandbox games, or the ideas for a house design and characters clothes. But i really hate anything that includes ai on art itself