A thread on how to recognize and avoid AI #gameassets! Using these game assets is not allowed when releasing a game on Steam, but might also lead to other legal issues as it's hard to determine who owns the rights. How to recognize them; 🧵
Generating assets using AI takes no effort, so these packs will often include hundreds or even thousands of similar assets. They can't judge what is good art, so they'll just include everything that gets generated
Take a look at any other game assets published by the same user. Often their art styles won't match. You'll often find wildly different styles, even in a single package!
AI users are often in it for the money, making some quick cash. They don't care for the craft or contributing to the community. That means they'll often sell a lot of packs for little money, a few cents or a dollar - anything but free
When they'll use AI for generating art they'll also often use AI to generate text. Their descriptions are overly specific and often explain the most basic things (like what 'pixel art' is) or use MANY words where only a few were needed
Sometimes they'll be honest about using AI, but often they'll try to justify the work they put in. Effortless stuff like resizing images,picking the best results or removing the background
Take a look at promotional images or thumbnails. They'll use random text elements using fonts that don't fit the aesthetic. An actual artist will also be able to create a decent looking preview, sample or thumbnail. They're proud of their work and want to show it off properly
When releasing pixel art they'll never include actual tiles, but rather complete buildings or sets. AI can't generate these tilesets and it takes a great deal of effort to make sure that tiles fit together and are modular. Something they do not want to invest time in
Using all of these tips will make it easier to recognize AI game assets! Stay safe, and happy holidays! 🎄🫶

@kenney what I hate the most is that these tools can be used to create decent graphics if one puts in the work and the individual graphics show that .

however the people making these asset packs don't put in effort to make the whole graphics packs good quality .

like , the moral background of how the models have been trained aside for a moment , I think that they could be used to make a good image pack . could I do it , at least to prove a point ? no ! I both don't have the skills required to get image transformers to output good looking images nor the knowledge to compose a good asset pack . these skills do impact the quality of the work while taking effort to develop and I think that should be acknowledged .

if someone , who not having an ability to create art themselves , makes a asset pack which is consistent , isn't just topic vomit and most importantly is good and usable in the context of creating a game , I don't see an issue with them selling it , as long as they disclose they used a text to image model .

that said the way the current generation of image transformers has been trained is not that ethical , though admittedly I am unsure on how ethical it could be , at least under the current socioeconomic system

I hate how all of this very interesting and powerful technology is taken by people who don't fully grasp its limitations to then exploit for financial gain while over blowing its capabilities

@kenney I'm terrible at thumbnails/samples/previews but would still like to consider myself an 'actual artist'. 😛
@StuffByBez Don't be too hard on yourself!
@kenney Crazy to see this...
I know on youtube appears channels with infinite-content with "ai-generated nonsense related to currently trending-context" - and those channels/videos get hundred of thousand of views.
Seems it happening everywhere - quantity always win.
@danil @kenney it’s been proven time and time again that the general population just doesn’t care enough about quality. Look at how wildly successful cheap stuff is… games and websites filled with awful ads, cheap fast foods etc… It’s a battle that can’t be won.
@cwagdev @danil @kenney
Had me until "fast food". There's a real means-based reason why fast food wins, and it's not because those people don't care about quality.
@kenney Thanks! This was very informative!
@kenney Another good indicator is simply the output, often posting new huge asset packs every single day.

@kenney The biggest giveaway is however can be known from how the AI works in simplified terms. It uses statistical probability to calculate what pixel comes after the next within a certain context window.

Image generation AI is just a very spicy Wave Collapse Function Algorithm, and knowing it makes sense out of all the mistakes it makes (weird fingers, background breakage, messed up patterns, etc.).

@PixelPerfectEngine @kenney I'm not sure if that is correct, although I'm not an AI export either… My understanding is that image generation AI works on the image in a global manner, removing noise everywhere at the same time in each iteration, so there's no "context window"…?

Otherwise wouldn't we be seeing errors like "infinitely long fingers" rather than just messed up structure (which can also be explained by the AI just ending up in some sort of search space local maximum)

@mao @kenney We never really saw into these algorithms (especially of big corporations' like StableDiffusion and Dall-E), so it pretty much could just do it on a neighboring pixel basis under the hood.
@PixelPerfectEngine @kenney isn't stable diffusion open source