420
"Steam’s required AI disclosure is the right move": Devs say gen AI should be called out in games and support Valve's store mandate (www.gamesradar.com)
submitted 3 months ago by to c/pcmasterrace@lemmy.world
37 comments fedilink hide all child comments
you are viewing a single comment's thread
view the rest of the comments
[–] -1 points 3 months ago (20 children)

How would you define use of AI assets?

Should using them in the prototyping phase of development require a tag? I don't want AI-generated assets in the final product, but I think I'm okay with them being used in test builds.

For me, as long as I can't spot its use in a publicly available release, then it doesn't really matter to me if it's used for either code or assets.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 20 child comments
    •
  • [–] 5 points 3 months ago (16 children)

    If AI assets were used at any point in development, there's no telling what else they cut corners on. If it's not immediately noticeable, that's worse, not better.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 16 child comments
    •
  • [–] 0 points 3 months ago (15 children)

    For prototyping you usually use any assets you have lying around, using AI for that is not cutting corners it's more expensive and labor intensive but gives you a better feel for what you're trying to build. Usually prototypes are done with either plain colors or just assets you downloaded from random sources or have from previous games, it's just to get an idea whether the gameplay is fun enough to build the entire thing.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 15 child comments
    •
  • [–] 0 points 3 months ago (14 children)

    So it's more expensive, wastes more time and fucks up the environment. That's not the flex you think it is.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 14 child comments
    •
  • [–] 2 points 3 months ago (11 children)

    Tell me you've never worked on a game without telling me you've never worked on a game.

    Sometimes (and by that I mean 99% of the times) you just don't have the funds or time to make assets for a prototype, you need a fast iteration cycle to nail the core mechanics. You need to learn as soon as possible if the game works or not, and while doable with Mario and Pikachu sprites it's much better if you can get a feeling for the atmosphere as well to see if it matches. This is the first hours of game development, after the initial GDD and before you even consider the full scope of the game.

    But I don't expect someone who sees in black and white to be able to understand the many scales of grey there are in this. Using assets without copyright for them is also wrong, yet that's a daily occurrence in prototyping, because you're not going to spend time and money building assets for an idea to check if it's worth investing time and money into, that defeats the purpose.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 11 child comments
    •
  • [–] -2 points 3 months ago (10 children)

    Using assets without copyright for them is also wrong

    Yeah, okay, we're done with this discussion. Pro AI and pro copyright? I pity whatever projects you were allowed to contribute to.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 10 child comments
    •
  • [–] 3 points 3 months ago (6 children)

    You're delusional. 99% of game makers care very much about copyright.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 6 child comments
    •
  • [–] 0 points 3 months ago (5 children)

    Pray tell, how can an indie game dev enforce copyright? The system is designed in favor of corporations with army of lawyers, specifically to fuck over smaller creators.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 5 child comments
    •
  • [–] 1 point 3 months ago (4 children)

    If it‘s so easy why don’t we see a myriad of copyright/trademark violations by large companies?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 4 child comments
    •
  • [–] 1 point 3 months ago (3 children)

    Because you aren't looking

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 3 child comments
    •
  • [–] 1 point 3 months ago (2 children)

    Can you share some examples?

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 2 points 3 months ago (2 children)

    I'm not pro AI, I'm fairly against it in most uses, but I do recognize it has some applications, and trying to close your eyes to it and day it's all bad is just not productive to any sort of serious discussion about it.

    As for copyright, yes, I'm in favor that some form of copyright should exist, although companies abuse that right and that shouldn't be allowed.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 0 points 3 months ago (1 child)

    It has some applications. None of them are in software development.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 3 months ago

    Well, first of all the majority of the discussion was around assets which are not what most people have in mind when they talk about software development. But also LLMs work great for their intended purpose of auto completion, even a local model running on your own machine is miles ahead of the snippet libraries we used to use until recently. Also they can spot errors in your code much better than you yourself can, as they're essentially impartial and looking for patterns that have previously been reported as causing issues statistically, while you are biased towards the mistakes you usually make.

    I wouldn't trust anything it outputs any more than what I would trust something written by an 8 year old on heavy doses of stimulant drugs, but in reduced scope and limited capacity they're quite useful, and can help in some tricky situations like memory corruption or threading issues.

    That being said I strongly support that vibe code is not trustworthy and anyone who does that is essentially gambling with that code.

  • permalink
  • fedilink
  • source
  • parent
  • [–] 4 points 3 months ago (2 children)

    I sorta agree with you in that I don't care about AI assets being used for prototyping, mood boards or similar. I think that that's the right way to use LLMs, because those are assets you wouldn't generate anyways, it would either be a solid color or more likely a random asset you grabbed online.

    However I don't think the argument should be that if I can't spot them it's fine, as you might not be able to spot some AI assets in the final build. I think that "no AI assets in the final product" is what I expect the no AI to mean, I don't see using AI as a stepping stone on things you wouldn't hire someone to do anyways problematic.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 2 child comments
    •
  • [–] 1 point 3 months ago (1 child)

    Just fyi, llms can not generate images. Llms call a diffusion image generation model as a tool to create images.

  • permalink
  • fedilink
  • source
  • parent
  • hideshow 1 child comment
    •
  • [–] 1 point 3 months ago

    I thought difussion algorithms were essentially Markov chains applied to images instead of text, so they would be an LLM with the language being images. But honestly never cared to look to deep into it so I'll take your word for it that they're different.

  • permalink
  • fedilink
  • source
  • parent