Star Realms maker Wise Wizard Games defends using AI art in new projects, board game artists call out ethics of decision

Star Realms maker Wise Wizard Games has become the latest well-known board game publisher to begin using AI generated images in its projects, despite widespread criticism from artists that the technology is built upon stolen work.

Wise Wizard’s use of AI generation in the artwork for its new trading card game, Draconis 8, is the latest instance of the controversial technology creeping into larger-scale tabletop publishing houses, following high-profile early adopters including Terraforming Mars maker Stronghold Games and Nemesis publisher Awaken Realms.

Despite condemnation from across the creative industries – more than 34,000 authors, actors, musicians and artists including Radiohead’s Thom Yorke and writer Kazuo Ishiguro have come together in recent weeks to call out AI technology as a threat to their livelihoods – use of AI art in board games seems to have had little impact to date on the success of projects that make use of it.

Board game publishers, artists and players took to social media and BoardGameGeek forums in September last year to condemn Stronghold Games using AI art in the development of the latest Terraforming Mars Kickstarter campaign. But that campaign still saw Stronghold raise more than $2.2m from over 19,000 backers.

Awaken Realms collected $3.8m for Grimcoven, which includes AI generated images, via a Gamefound campaign this summer, and games such as the $1.2m Wonders of the First CCG have actively leaned in to the use of AI artwork as a selling point.

Large board game publishers to actively come out against AI artwork in a meaningful way, meanwhile, have been few and far between.

The highest profile of those so far has been Wingspan and Scythe publisher Stonemaier Games, with co-founder Jamey Stegmaier telling BoardGameWire in April that his company wants “nothing to do with it”.

In March this year Awaken Realms had to take down AI-generated promotional images for its Puerto Rico 1897 Special Edition crowdfunding campaign, after being contacted by the game’s licensor Ravensburger and reminded that generative AI “cannot be used in any part of the art process”.

This summer Spiel Essen – the world’s biggest board game fair – stopped using AI generated images to market the event, citing criticism from industry professionals and consumers about last year’s campaign and ongoing uncertainties about the legality of the technology.

Wise Wizard raised more than $120,000 for its Draconis 8 Kickstarter last month as part of an inventive push into the TCG space, with the small-footprint game based around $5 packs of complete, eight-card decks.

But that push has also seen the company lean heavily into using AI algorithms to generate individual images for the large number of cards in the game – which the company says would not have been possible without the technology.

Wise Wizard CEO Rob Dougherty and Draconis 8 lead artist Antonis Papantoniou have responded in length to questions put to them on the subject by BoardGameWire, and the entire Q&A is included below this article.

In it, the pair attempt to draw a distinction between a “non artist” producing AI art via prompts, and their own method, in which a professional artist feeds their own hand-drawn prompts into programs such as Dall-E or Midjourney, and paints over parts of what it generates.

They add that without using AI, the game would only have 17 pieces of art, compared to the 321 they are able to include through AI image generation.

BoardGameWire also reached out to high-profile tabletop gaming artists about the situation – several of whom replied describing AI art as theft, and criticising Wise Wizard’s decision to engage with a technology that carries heavy ethical and environmental concerns.

In an extensive response which can be read in its entirety here, Root artist Kyle Ferrin decried AI art as equivalent to mixing slop for pigs, castigated the technology for its reliance on stolen art and said it was “embarrassing” that the industry was even having to have a conversation on the subject.

Undaunted and Final Girl artist Roland MacDonald told BoardGameWire, “It is sad and disappointing but unsurprising. Companies having no morals about using tools built on stolen art is one thing, but for an artist to do it is ugly. In the end despite the artist efforts it just looks like generic AI chaff.

“The work is visually bold but lacks any personality or charm. Maybe that is what they were going for. Hopefully it will just be a fad and the market will get bored of this soon.”

Joan Guardiet, who has illustrated games including The White Castle and Finca, told BoardGameWire, “Having read about the creation process detailed by the artists involved in the project, I can see how this could be an interesting way to work for some specific individuals.

“However, there is still the question of all the ethical ramifications that can’t be overlooked. As it is designed right now, AI and the companies using it profit from the work of thousands of artists that have not been compensated or acknowledged in any shape or form. Nor can we forget all the unnecessary environmental costs of AI.

“I’m not against AI as a concept personally, and this detailed process is an example of something I could be interested in trying, but until all the legal, ethical and environmental aspects have been resolved I’m not going to use it or support it.”

Q&A with Wise Wizard CEO Rob Dougherty

Rob Dougherty: Before I get into the questions, I have to say something about the terms “AI Art” and “AI generated images”. When you hear those terms, you think of a non-artist typing in a prompt on an AI art tool, then cutting and pasting the resulting raw image “art” into their final product.

That is NOT what we are doing with Draconis 8. In Draconis 8 we use AI as an artist’s tool, not an artist replacement. In our process, there are hand-painted prompts and many points along the way where the human artist paints over the AI’s output. We lay out the full details on the Draconis 8 Kickstarter page.

So it is fine to use the terms “AI Art” and “AI generated images” when talking about Draconis 8, but I want people to know what we are talking about in the context of Draconis 8.

BoardGameWire: Is this Wise Wizard’s first use of AI generated images in a game, or have you used the technology in other projects – either for internal art, or box covers, marketing etc?

This is our first major project using AI as an artist tool. Our digital team has used AI art as placeholder art during testing, and we have used a similar process of using AI as an artist tool to create some pieces of art used in our Hero Realms digital game.

Where did the idea come from to use AI generated images for Draconis 8? Was this a business-focused decision that you brought to lead artist Antonis Papantoniou, or was it suggested to you by Antonis – or something else?

It was a collaborative decision to use AI in the art processes for Draconis 8. Antonis loves to play around with technology in the creation of art. He was independently looking at AI tools and talking to us about ways they could be used. I gave him the green light to use some of his work time to explore techniques.

Draconis 8’s designers used AI to make art for their prototypes. This is a pretty common thing for game designers to do with a prototype, just to make something quick. It doesn’t mean they intended to use AI in the final product. There was no artist in their process, the art didn’t look great, but it was good enough to get their vision of having dynamically created unique cards with different art and stats.

When I played Draconis 8 for the first time, I loved the game play and potential for the product. I talked to Antonis about the possibility of using some of the AI techniques he had been experimenting with in his process to try to capture that unique card feeling, but with a higher art quality.

We initially created a model with art layers where the art would be dynamically generated based on the stats on the card, i.e. subject of the piece, the background, the weapon being held, etc would be modified depending on the stats on the card. While this “worked” technically, the resulting artwork didn’t have the great “look” typical of Antonis’ work.

We eventually decided to alter art based only on the Speed and Ability of a card. Antonis could now fully compose each art piece (with beautiful results), and make the art pieces extremely thematic, for example the Fear ability cards use undead art. Each deck has 8 cards, and this model resulted in 320 total pieces of art, which means most decks won’t see any duplicate art.

For example, in this card spread, you can see that these three cards all have the Aura ability, but each have a different speed and the corresponding piece of art for that Speed combined with the Aura ability.

Kickstarter’s ‘Use of AI’ section for campaigns has received criticism for allowing creators to provide quite vague details about their use of AI generated images. But you’ve included a separate 500 word section for Draconis 8 detailing how Antonis used the technology – why did you decide to do that?

Mainly we wanted to be super transparent about our use of AI. AI is powerful, scary, and has incredible potential to do good or harm. It can be used poorly, like replacing a human artist and filling our world with junky uncanny valley garbage “art”.

We think we are using it well. As an artist’s tool that assists their creative process, AI removes some of the drudgery, and allows an artist to more easily create beautiful works of art. We wanted people to see what we are doing and decide for themselves if they think it is an appropriate use of the technology.

Are you aware of the main criticisms from artists of the use of AI generated images – that artists are not consulted or recompensed when their work is scraped for use by companies making AI tools, and that using AI images takes work away from human artists? What’s your position on both of those arguments?

These are very complex ethical questions. I have my point of view, but I understand and respect that others may look at things differently. That said, here’s how I look at it.

Throughout time, from the first cave painting, to the art of ancient cultures, to medieval art, to the art of renaissance masters, to the modern art classroom, we have looked at each other’s work and copied each other’s techniques. There is an inherent understanding that if you create a piece of art and put it out into the world, other artists will look at, and learn from it.

Humans have now created machines that look at and learn from basically every piece of artwork on the internet. To me, this doesn’t seem like stealing, it seems like the same thing we have always done. Looking at and learning from art that is in the world. However, there are examples of people using AI tools in a way to intentionally recreate copyrighted art or art in a similar style of an existing artist, and I do not think that is an ethical use of AI tools.

In Draconis 8, Antonis, Wise Wizard Games’ full time employed lead artist made the art. The use of AI as a tool didn’t replace him, it allowed him to create a large set of gorgeous, cohesive art according to his personal vision and style. Without the use of AI, the game likely would have had 17 pieces of art. One for each ability, one for each color of dragon, and one terrain art. Using AI, Antonis was able to do 20 pieces of art for each ability and color of dragons, creating a rich and diverse fantasy world.

One point of ongoing contention with AI generated images is that the algorithms can only create images by being fed large quantities of existing artworks, largely without the consent of the original artists. The Kickstarter AI terms asks if you have the consent of the owners of works used to produce the AI generated parts of the project, to which you’ve replied yes. But you don’t have the consent of the artists that have had work fed to Midjourney and Dall E to create the images for you – shouldn’t you have answered ‘no’ to this question?

We have the consent of the artist on our team, and we did not feed any images or prompts from other artists into the model. It was hard to understand exactly what Kickstarter was asking here. Do they mean the full creation details of every tool used? Our process of using those tools? I tried to give a very detailed answer that would give the reader the information they needed.

“Yes. This art is created by our lead artist (a full-time employee) using the technique described above. No other artist’s name or work is used as a prompt. AI tools used include Midjourney, Dali E, Photoshop AI Tools, and a local installation of Stable Diffusion and ControlNet.”

We listed the tools used, so if you don’t like how those tools were created, you know we used them. Again, we were going for as much transparency as possible.

Did you expect any backlash over using AI images in Draconis 8 – from professionals within the industry, and from gamers? What’s been the reality of the situation?

Yes, we expected some backlash, and there have been some negative reactions. We engage, explain the process, and invite people to read about it on the Kickstarter page. While not everyone agrees with our perspective, many people like the way we are using the technology as an artist’s tool.

Could you have made this game without using AI generated images? How many more artists would that have required – or how much extra time would it have taken?

Yes. If we didn’t use AI as an artist tool, we just would have had far less art in the game. Without the use of AI, the game likely would have had 17 pieces of art. One for each ability, one for each color of dragon, and one terrain art. Using AI, Antonis was able to do 321 pieces of art, including 20 pieces of art for each ability and color of dragons, and a terrain art, creating a rich and diverse fantasy world.

What do you estimate the cost saving has been on this project in using AI generated images, and how has that translated to the cost of the final game?

Without the use of AI, we would not have been able to invest in such a large number of pieces of high quality art for a new game, so we would have created the game with significantly fewer art pieces.

Are you planning to use AI generated images in any of your other upcoming games? And has the technology changed your strategy for 2025 – can we expect more games / expansions from Wise Wizard than you initially planned, or games with much greater amounts of artwork?

We have plans for multiple games where the art is created without AI technology by an independent artist, we have upcoming games where Antonis is creating art without AI technology (like the next big Star Realms set), and we have upcoming games where Antonis will use AI as part of his process. Using AI as an artist’s tool on some projects will allow us to make more games with more art, while keeping that art at a very high standard.

Are you worried about potential copyright claims from artists? Given that Dall-E and Midjourney have both been known to reproduce copyrighted materials. What happens if it’s recreated another artist’s dragon for you?

Antonis is painting prompts, and painting over the output from the AI again and again. Every pixel has been hand painted and/or hand painted over. That process can’t result in some other artist’s copyrighted material.

What would you say to artists out there looking for work in the board game industry – do they need to be using AI tools to survive?

When I started making games and commissioning artwork, that artwork was painted with oil paints on canvas. The resulting physical painting was then photographed or scanned to get a digital image used for production. Later, some artists started making art directly on their computer using a digital paint brush. Some wondered if that really counted as art. Where was the original?

Now, most art is made using a digital brush, but some artists still paint old school without a computer. Their work takes longer, revisions are harder, and they have to charge more. That is fine. In the same way, some artists will use AI as a tool and some won’t. We don’t believe that artists have to use AI tools to continue to be successful in this industry.

Q&A with Wise Wizard lead artist Antonis Papantoniou

BoardGameWire: Wise Wizard told us regarding your previous work for them: “He has created many of our best pieces for our other games without the use of AI”. Why did you decide AI art was the way to go for this project?

This project called for a wide variety of images within a tight time frame, and a restricted budget. It also called for a large amount of dragon artworks which could eventually become repetitive, possibly resulting in stale images. To solve this, I combined my drawings and paintings with AI tools for generating guided variations of my compositions and of specific elements. This allowed experimentation and exploration of options that would not be otherwise practical under the constraints of deadlines and budget. It also made the image-making process more creative and adventurous.

Have you used AI generated images for any of your other previous board game work? What did you do, and why?

Although I used AI tools for upscaling, patching , color adjusting or extending images in some of my previous work, Draconis 8 is the first project where I used my sketches as AI prompts, combining image renderings and paint-overs to create AI-assisted works.

For Draconis 8, when you add your rough image prompt into the AI image generator and ask it to create options – it’s then using other artists’ work to generate the new art for you. How do you feel about the uncredited artists’ work being used in that way?

As a game artist, I understand that this is a complex issue. I think that AI tools can be very useful in boosting creativity and speed, but the usage of AI generators to produce raw images does raise ethical concerns. However, my process when using AI tools is focused on creatively interacting with the AI, rather than relying on raw AI results.

The image-making process in Draconis 8 begins with color sketches of the main concept which are speed-painted by me. I load these into an AI engine and use them as guidelines to explore variants, which I further control and guide with my rough lineart thumbnails.

My textual prompting directions are very simple and generic, with no references to artists, styles or specific artworks, so that the AI generation can focus closely on my sketches. I then combine parts and pieces from a variety of results to create a composition. The composition is then painted-over by me and loaded back to the AI for quick partial renderings. The renderings are again combined together and painted over, repeating this cycle of back and forth multiple times to progressively reach a final image. I feel that this continuous artistic hands-on involvement allows me to creatively control the final image and style, which is very different from a direct raw AI image generation.

How long did it take to create the more than 320 images for the game, using the AI tools as you did? And how long would it have taken to create those artworks without the tools, to the same standard?

The average pace for this project was about 1-2 images per day. Without the use of AI assisted tools, including renderers, upscalers, patch tools and adjustment tools, each image would take about 2-3 days.

Are you aware of the main criticisms from artists of the use of AI generated images – that artists are not consulted or recompensed when their work is scraped for use by companies making AI tools, and that using AI images takes work away from human artists? What’s your position on both of those arguments?

I believe that artists should be recognized and compensated for their work. I have worked for many years as a game artist, so I am sure that my own images have been used in AI training datasets without my consent. It would be ideal for the AI research and deployment companies to support artists, compensating and crediting their work and there should be at least options for artists to opt-out from AI training datasets.

Are you planning on using AI art for any of your other upcoming projects? And have you any other new ideas about how you might be able to use the technology?

Ι find AI tools to be extremely helpful in speeding-up the creative process. But it also depends on the way the tools are used. I believe that an interactive process of creative back and forth with these tools can yield satisfying results. I would love to use AI tools in future projects and explore more ways of combining my traditional drawing, painting and sculpting skills with new AI technologies

What would you say to artists out there looking for work in the board game industry – do they need to be using AI tools to survive?

As I have already mentioned, AI tools are very useful and I would add them to my own toolkit as a productivity and exploration assist. However, I don’t think that these tools are necessary to survive or succeed in the game industry. To use these or any art tools effectively, you will need to have a solid foundation of aesthetics, design and visual development that transcend medium and technology. Therefore, my recommendation would be to focus on a deep understanding of the core aspects of visual communication which include form, value and composition, applying them to visual storytelling and on enhancing the gameplay experience. When creating game art, the use of art and design to enhance player interaction should be the primary goal, and that is, in my opinion, the key to success in this field.