Skip to Content
 

State of AI Generated Art in Board Games Survey.

12 replies [Last post]
Jacob
Jacob's picture
Offline
Joined: 03/04/2024

The AI controversy has been an upcoming debate in recent years. This survey wishes to explore the effects that the debate is having on the purchasers of board games. I would like to build a resource for others to help them understand what is the general opinion of hobby board gamers around the world on this issue. If you would consider taking 5 minutes to fill out this survey, your time would be greatly appreciated. You can also subscribe to the BGG post to see the final results once the survey closes. Thank you so much and I look forward to your responses!

Survey Link: https://forms.gle/cL79WuAqLCLMvjHV8

BGG Link: https://boardgamegeek.com/thread/3536844/state-of-ai-generated-art-in-bo...

X3M
X3M's picture
Offline
Joined: 10/28/2013
Filled in the survey

The line " AI art usage is disclosed before purchase ".
Is very important here.

I would love to see something along the lines of " AI art is NOT used, is disclosed before purchase " as well.

Personally, there is good AI art. But I hate it when it is not obvious anymore. On several social media, you can find these very real looking, AI generated video's.

And another argument, why I rather see real art. Artists are going to have a very difficult time. I even prefer some bad art over an AI generated image. They work for it harder than someone who simply invests in cheap AI generating programs. Developed skills for it. It all goes to waste. AND...AI generated images are based on images of artists. So, the picture is actually plagiarized from several sources.

Also, in personal experience. Certain people could be faking development. That is all I say about this nasty subject.

AI art as placeholder...? Well, this goes in a grey area. I do it myself too. But I will never sell AI art.

There is another grey area. What if an artist only allows art it made itself. And allow the AI to toy with it? I have yet to see this.

Good survey. The questions are well addressed.

Cheers, X3M

questccg
questccg's picture
Offline
Joined: 04/16/2011
My opinion is a bit different...

And it comes out when you ask questions like a "small company" (< $250,000 annual income) well it matters more when the company is BIG and has the funds to make a project with investment into art (Not AI generated).

I think this is a very important distinction. Why?

Because if you know that you are RUNNING a "small company" you know a bit about the finances behind "Designing Games". You're either well to do to start with like some money put aside to help deliver your "creations" or you have very little money to start with.

Where it gets a bit fuzzy is when companies with more income (> $250,000 annual income) don't DISCLOSE that they have used AI Generated Art is where it gets me. Because this means that a company who can AFFORD to pay an illustrator chooses NOT to do so.

I love working with Artist and I've worked already with three (3) different ones. But the reality is that my "initial fund" is nearly gone and the reality is that my games need to START SELLING if I am going to continue down this very difficult path known as "Game Design".

Again why? Because it's a very competitive market and very niche in terms of what moves forwards and what does not. I'll be the first to disclose that "Battle Botz" (BB), my robot battling Card Game is using AI Generated Art. And without the AI Art, I would not be able to create illustration that look so ROBOTIC in nature and it could be difficult for an artist to render such creations as well. BB is still in early design phase and I have shelved the project just a bit... While I focus on SALES to try to earn some income.

Independent of the budget I would have, BB is an AI Generated Art Project. I want it to be that way simply because the robotic art is HARD to reproduce.

And I'm making like $0 annual income from designing games too.

I feel if you can't find an artist to make the kind of ART you WANT for your "design", well if AI Generated Art works for you... I would therefore use it and be done with it. If however you are a large company with a BIG budget (like > $250,000 annual income) well them I would only disclose that AI Art was used in a certain way (perhaps only one aspect of the game) and also be done with that too.

But having the money to pay of an Illustrator does NOT make a project go. Like I said there are situations in which AI Art is better than human made art and in some very niche applications (like my BB) it's hard to design the concepts that are required for the game/design.

So there are projects that CAN and CANNOT benefit from Generated AI Art.

If it's a question of NOT BEING ABLE to bring a product to market because you have too little INCOME or finances to pay for a real Artist or Illustrator or the subject matter (Topic) is specialized such that real art is hard to come by... Well then these are questions that remain debatable.

I know many will NOT like my position on this matter. Like I said, I enjoy working with Artists and Illustrators... But only when it is FEASIBLE. If it is NOT feasible (for whatever reasons and I've covered some of them above), well I believe that AI Generated Art is acceptable (as an alternative).

Your projects should not suffer just because you don't have sufficient INCOME or don't have someone who is capable of drawing what you are looking to bring to market, etc. etc.

I await the results of your poll... And see who agrees with whom.

Cheers!

questccg
questccg's picture
Offline
Joined: 04/16/2011
What is SMALL and what is BIG???

I guess one of the "core" pillars to this debate is also what is SMALL and what is BIG. IMHO I think > $100,000 is BIG enough to afford a real Artist or Illustrator given that there are no restrictions to the type of Art you want to produce. Like I said "Battle Botz" (BB) would NOT be possible without AI Art and that's due to the nature of the subject matter.

If you have > $100,000 in terms of annual income ... Well then I believe you should set some of that aside for when you WANT to make a game and bring it to market.

I've done this with "Quest Adventure Cards(tm)" (Quest) and about 60 or so pieces with a professional Artist/Illustrator came to $5,000 CAD in cost to produce. TradeWorlds (TW) was a bit different, it cost about $8,000 CAD in cost due to the high quality of the rendered Art ... For Quest I have yet to see any income, and for TW I broken even and recouped my $3,000 CAD investment into art. The plain fact of the matter was that with TW I had a Publisher and they covered the remaining art which was about $5,000 CAD... And paid me back the $3,000 CAD I had initially invested.

So I believe it depends on the availability of income when you decide to launch yourself into designing Board and Card Games. If you are working and this is a HOBBY... Use Generative AI Art for your own prototypes. Spend a couple bucks and have yourself something better looking.

But if you are SERIOUS and want to bring your game to market... Look at your disposable income and your salary and figure if you can spend 5% to 10% on Art for your project. If YES... Well you'll have a NICE game that you can see if you can find a Publisher or Self-Publish via a Crowdfunding Platform like Kickstarter, Backer Kit or Pledge Manager, etc. etc.

Remember you are taking a RISK. And the millions of dollars you are going to make are only in your mind. Because of the nature of the competitiveness in designing Board and Card Games... It is not impossible, just highly unlikely!

questccg
questccg's picture
Offline
Joined: 04/16/2011
What to DO and what NOT to DO... A critical analysis in failure

So had I been a bit more patient, I would have learnt that MY approach to designing and manufacturing of "Quest Adventure Cards(tm)" (Quest) was incorrect.

I had put aside a fund of $30,000 CAD after several years of consulting with good sales numbers and sufficient EXTRA income for my "war chest".

Where I went WRONG was to decide to MANUFACTURE 10,000 Booster Packs for Quest. It was a colossal loss of $10,000 CAD to make the cards and $5,000 CAD to package them in booster packs (since the manufacturer did NOT have the capabilities to do BOTH).

Then as mentioned above $5,000 CAD went to Art and I was left with about $10,000 CAD after all was said and done. So my total LOSS was about $20,000 CAD after all was said and done.

My mistake (and it's important to understand) was that I MADE the game before I could ESTABLISH a RELATIONSHIP with SOMEONE who could SELL it. That sentence is very important and let me explain.

questccg wrote:
If you want to make a GAME (Board or Card) and you have some money set aside for it... Here is what you should do!

#1> Make a quick and dirty prototype.

#2> Playtest with various groups and develop it further.

#3> Once you are ready... Try to produce SOME ART given an artist and launch a Crowdfunding Campaign.

#4> If you succeed... You should have enough money to complete the Art. Make sure you account for this COST when determining your Funding Goal and the costs to SHIP and pay 10% as fees for the Crowdfunding Platform.

#5> If you fail your Crowdfunding... All is not all lost. But the road is much more complicated...

#6> Try to find someone you can SHARE your passion in designing with whom you can partner with. Maybe they can help with SALES and help get you a deal in terms of some kind of agreement to sell.

#7> If all fails, you have a NICE game that you can choose to either FINISH with more art... Or shelve the design and never design another game again!

***

Okay the last item (#7) is a bit HARSH ... But it is TRUE. The road after a Crowdfunding Campaign is VERY TOUGH. Unless you know SOMEONE... Or have great networking skills that allows you to FIND the right person. It's going to be a loss in most instances.

BUT REMEMBER... A loss of $2,500 is much less than a loss of $20,000 CAD... You cut your losses early and spent a little bit and have not lost all your precious income. You tried and it maybe goes your way and maybe it doesn't.

Either way... DO NOT MANUFACTURE before you've established a SALES channel. Or a way to bring your game to market.

I learnt my lesson... It was costly ... But at least Quest is NOT dead... There is still some potential yet to be discovered soon enough!

Best.

questccg
questccg's picture
Offline
Joined: 04/16/2011
Had I not manufactured Quest...

I would have only "lost" $5,000 CAD for the making of all the Artwork. But even there it would NOT be a LOSS because I could re-use the art for a re-launch or a 2nd Edition which would prove to be more interesting and in the proper format.

This is the "extra" notice to point #7 which said something like: "You can decide to complete the art for your game or shelve the design..."

So in my instance, I could have finished the ART and the total cost to me would have been a fraction of the LOSS that I had taken (+$15,000 CAD more).

The bottom line is this: try to minimize your LOSSES and maximize your chances for success. Crowdfunding is good and if you don't succeed the first time, I have seen people do up to three (3) Crowdfunding campaigns about the same game.

Why and HOW(?) exactly???

Well it stems from the fact that you make the campaigns more and more exciting as you perfect your end-product. Same goes with changes to the campaign to make it more attractive to potential buyers.

So you've got three (3) tries at it... Focus on the positive each time and ensure that you get to re-contact people who were Backers in a previous version of the campaign too! That gives you a pool of Backers to start with and hopefully your campaign becomes successful given that each successive iteration meant new potential Backers.

Sincerely.

questccg
questccg's picture
Offline
Joined: 04/16/2011
@Jacob good poll and/or thread

Be sure to post the results or a LINK to BGG if you prefer. I rarely get the CHANCE to voice my opinion on BGG before someone tells me "You are wrong!" And that's usually the NORM over there and then more people criticize my Point-Of-View (POV).

This is the first time that I have seen people on BOTH (and in the middle) SIDES of the argument.

My POV is simple: if the AI has learnt how to DRAW and no longer requires images to make more illustrations and furthermore if said data can reside on a Home PC (like "Stable Diffusion" and their various models) well then it's a bit like me browsing the web for images of the American Flag. Since I don't know how many White & Red Lines in the flag, I see what is available on iStock for $x.00. I then decide with the information that I have LEARNT and use that KNOWLEDGE myself in illustrator and draw me a US Flag. Is this ILLEGAL? Is this IMMORAL?? Knowing full well that the person who drew the original Flag will NOT get any ROYALTIES because I made my own with SAID INFORMATION.

If that is wrong... We have a real problem with society. Or some people don't understand the implications of the side they are on.

Unfortunately I don't have the details as to HOW the AI goes about making Art.

And I THINK that if AI creates it's OWN "interpretation" of a prompt (whatever Technology you use...) then that AI should not only be LEGAL but be able to be used. If it's the SAME as me googling for a representation and then making MY own artwork, AI Generative Art is an innovation and it is NOT copying the source images used to TRAIN it.

If a AI Generative piece of Software like "Stable Diffusion" which *offers* several types of Models for different types of illustrations (like Comic, CGI/Rendered, Photo-realism, etc. etc.) can fit on a personal computer (because you must download the software and run it locally).

This means this is not some kind of issue with Bitcoin which requires a bunch of servers all trying to solve a mathematical equation and wasting electricity basically doing nothing important besides determining which "farm" solves the equation and earn the reward (mining).

This is obviously not the case with AI Generative Art. It's a low-overhead Technology which requires limited amounts of space to "generate" images. If an image takes 1000kbytes or 1 mb. The data it learns from such an image must be significantly smaller because nobody would be able to STORE such data on a home PC. And then you could argue ... Server farms cost a lot of money, require a lot of electricity and are environmentally "questionable". But this doesn't seem like the case with AI Generated Art... So that really makes me go... Hmmm.

Good stuff! I'm still in the middle. And I appreciate you taking the time to make that POLL and go through the outcomes (as you see fit)!

Cheers.

lewpuls
lewpuls's picture
Offline
Joined: 04/04/2009
Danger

Using AI generated art is dangerous because some "generators" include copyrighted art in their databases without permission of the copyright holder. Then the "generated" art may still be subject to copyright and to damages for use without permission. Just Say No.

questccg
questccg's picture
Offline
Joined: 04/16/2011
It's a bit more complicated than that TBH...

So the AI Generator was training on "copyright" images... However the images in question were never banned from being used for "Machine Learning" purposes.

Prior to the inception of AI Generated Art, "fair-use" by Human Learning was acceptable to a certain degree. There were copyrights on "Brands" like Disney Characters and other popular brands like Marvel, etc. etc. which meant that some likeness would be protected and therefore Copyrighted.

However as an Artist, you may inspire yourself from said "brands" and generate your OWN art and you would be still in "fair-use".

Furthermore my example of the American Flag, demonstrates that Human Learning is possible and is "fair-use" too.

So the argument is: this is a Technological Advance (Transformative), it doesn't require any further artwork to generate more artwork, and it used Machine Learning at the time of training.

Therefore the logical conclusion must be that AI Generated Art is a perfectly acceptable source of art for SOME cases... Let me explain.

If you are trying to have UNIQUE style across a Game, then the art may be hard to ensure that the various art assets have the same "Look & Feel" and even a simple LOGO can also sometimes challenging to achieve consistency. Some AI Art takes too many liberties and adds a sixth finger or a third eye, etc. etc. It is far from perfect.

My POV is ... Take it with a grain of salt: you can use it some instances but just don't expect to have ONLY AI Generative Art. In other words, don't bet the farm only on AI Generated Art. But also don't be surprised if you start seeing more and more of it! This doesn't feel like a FAD TBH.

Sincerely.

questccg
questccg's picture
Offline
Joined: 04/16/2011
It's a bit sad... but I expected it

From my read-thru the results, it seems like 50% of people don't want ANYTHING to do with games that have AI Generated Art. Be it small companies (under $250k) or companies that disclose the usage of AI Generated Art... Nothing really seems to matter, any one can see... Nothing really matters to THEM!

Which is unfortunate. I did learn more than I expected about AI Generate Art and how different platforms can even run on a Home PC. To understanding the number of parameters which go from 7 billion to 70 billion which make substantial quality improvements to the art but require more computer power.

I learnt the whole NOISE and process of decoding or encoding images to and from NOISE. All fascinating stuff.

Anyhow ... You've made the results public and that's good too. If you feel like decoding the answers that's up to you. But making it accessible to all is the easiest method TBH. And that's how it is currently.

You may get a different result from your own analysis... Mine own is pretty annoying the people who don't want EVs are the same people that don't want AI Art. I'm not sure how much you can REALLY do with AI Generated Art. Like in the latest BGG comment, I see nice art but can the machine pump out near identical characters to suit a "game"?! That IDK.

Can't wait to hear from you!

Jacob
Jacob's picture
Offline
Joined: 03/04/2024
Been working on it…

I have been a bit busy trying to get ready for the launch of my board game, but I hope to have some analytics by the end of next week.

larienna
larienna's picture
Offline
Joined: 07/28/2008
AI are trained on copyrighted

AI are trained on copyrighted images, but like I said in the other thread on BGG, good luck proving that the copyrighted content was used in the creation of an image.

It's like if I stole your eggs and made a cake with it. Even if you have access to the cake, hard to prove it was your eggs.

As I explained on BGG, with confirmation from ChatGPT, pixel art (low color) and chip tune music might be hard or even impossible to generate with AI. The lack of granularity makes the gradient impossible to follow, since the problem is too much discrete.

So I guess, going back old school is a good thing for maybe the first time in history. I wonder if the same phenomenon could also be applied to other part in life that could be threaten by AI.

questccg
questccg's picture
Offline
Joined: 04/16/2011
It's more about the drawing of the US Flag...

questccg wrote:
Let's say that I want an image of the American Flag. I google and find an image on iStock. I click on the link for the image and observe the number of Red & White Stripes. And there are 50 Stars. Rather than purchasing this art asset, I open up Illustrator and design my own with said information that I learned from iStock.

Is this wrong? It's Human Learning and it is not Copyright. What does this mean then? It means that iStock and the person who created the American Flag that was being sold got $0.00 in sales and royalties. And I got the KNOWLEDGE about HOW to make the American Flag.

Simply put I didn't STEAL anything, I just LEARNT from the image online. Same goes for the Machine Learning: they didn't steal the images, they just LEARNT from them. They don't need said images either to draw more art. Just as I don't need to BUY the American Flag on iStock.

Now there existed provisions in the LAW which allowed "Human Learning" but there was nothing about "Machine Learning". Obviously it's a bit different in scale and scope. And while as Humans we've been doing this for a LONG TIME, there is little in the LAW about "Machines" and how they can learn. Basically it's NOT new but this kind of Global application IS NEW.

Well relatively new (like maybe 5-Years). And therefore the LEGAL ramifications have yet to be "solidified".

Syndicate content


forum | by Dr. Radut