If AI art isnāt for you, I totally get it. No pressure. No agenda. I just wanted to give some context, because Iāve seen a lot of misunderstandings about how it actually works and what is involved in using itāespecially in projects like this.
So if youāre interested, hereās why I used AI, how I used it, and where I stand on the bigger conversation. I write about it because it's always the moment that people write off or dismiss the tarot deck, and knowing what was involved and what went into it, I don't have the same reaction.
I thought a lot before deciding to use AI. Iāve been making art since I was a kid. I worked as a designer for years. I knew I wanted full control over the visuals in this deck.
And for me, AI was about making this deck better than it would be without it. Part of that is that it doesn't just make things easierābut made quality more possible.
When I create traditionally, it takes forever to get something right. If I donāt nail it, I can't afford to start over. AI let me iterate faster, push my creative vision further, and refine every single detailāwithout being locked into one version.
And yesāsome people just type in a prompt, get an image, and call it a day. Thatās not what I did.
I spent months of full-time work experimenting, training, and reworking these images. Every single piece was a collage of multiple AI-generated elements, layered and refined in Photoshop. The goal wasnāt just to create something visually coolāit was to make sure each card felt right. I think a lot of people assume that it's a one-and-done thing. You write the prompt and you get the image, but it's still hours and sometimes days of photoshop work for each card after many iterations and feedback cycles, which is from years of perfecting that skill as a designer.
That said, if something came out that made it so I could spend a fraction as much time photoshop and collaging it, I would not stick by my old methods out of principle just to cherish how skilled I am. I would happily adapt. For me, the skill is a means to an end to create what I envision. It, in itself, is not the goal. If I could adapt and try out 30 different versions instead of being exhausted by trying to produce a single version, I would be thrilled and I think the product would be better for it. I care that artwork feels right and communicates the right thing, and communicates what I want it to communicate. I don't punish people for it taking a short or long amount of time. If it doesn't feel right, then that's its own issue.
Take Doomscrolling (4 of Feels). The original 4 of Cups? Some guy sitting under a tree, ignoring a golden chalice. Our version? A person floating in the dark cosmos, phone in one hand, laptop in the other, completely detached. Because thatās what disconnection looks like today.
AI didnāt replace the creative processāit expanded it. It started with a vision in my head, and then trying over and over again to achieve it, and getting something more magical frankly, than I think would have been created by my hands and that fit the aesthetic I wanted, which I don't' feel it would have happened if I had put it in someone else's hands.
Without AI, this deck would not exist with hand made artwork. It would not exist at all. For many, that's how it should be. For others, that feels like a shame.
ā No, I didnāt use other artistsā work as an input.
ā No, I didnāt use artist names as prompts.
ā
Yes, I used my own images, evolving them through AI to refine my style.
People assume AI just pulls from a database of stolen images and pastes them together. Thatās not how it works. It generates new images based on billions of referencesājust like how a human artist learns from everything they see, whether they realize it or not.
And honestly? When I make art by hand, I take way more direct influence from other artists. If Iām painting, Iām referencing specific works, pulling ideas, remixing compositions. With AI, the process felt more original, because it was about refining and shaping an evolving concept instead of copying anything directly.
And the biggest irony? I kept trying to make the AI-generated art look more like my style. In the future, I think this will become more and more possible.
I know the next question: But isnāt AI bad for the environment?
At the time IĀ made the artwork for this deck, no one was talking about this, and I wasn't aware of any implications but it. Since hearing all of this, I went and checked for myself. What IĀ learned is that training AI models? Yes, that takes a ton of energy. But generating images? Much less.
Creating these images used as much water as a 15-minute showerāa lot less than what Iād need to create 78 physical paintings.
To put that in perspective: If I had painted this deck traditionally, it would have used significantly more resources (from canvas, cotton, wood, and oil paint production). So no, AI isnāt good for the environment, but traditional art isnāt always the sustainable alternative people assume, either. The large impact comes from training these models, which you could argue my use encourages more of, but my use was not a direct cause of either.
Some of the new models coming out also use significantly fewer resources than the current models, and these will continue to improve (I also used to work in a data center while working at Apple and traveled around the country to the different data centers that store all of your photos, apps, and services that you use on your iPhone and know that they use a lot of passive technologies to mitigate these expenses). AI training is more resource intensive than your typical service but I believe AI will also lead to breakthroughs in the discovery and speed of implementation of novel and more efficient energy sources.
That said, it's worth remembering that early photography used chemicals like mercury, silver nitrate, and cyanide, which were toxic and polluting. The darkroom process generated hazardous waste that was difficult to dispose of safely. There is a question whether we've learned better and if situations are more dire now. It may be true. Would we have sacrificed the technology of photographs and gone on a timeline without it because of these ramifications? I'm skeptical. I don't think we'd have most of the technology we have today if photographs were not a bridge to get us there.
Further, when books became mainstream, they began deforesting large swathes of area to have wood to turn into pulp for printing. As much as 14% of deforestation was to service paper goods. I think books are important and worth having. [Industrial Environmental Performance Metrics, Challenges and Opportunities ā The Pulp and Paper Industry (1999)]
So we get into the question of whether the product of what we're currently making is worth the devastation it creates and if now is different because we know better and the situation is more dire. I think there is truth to that, but I also don't think even in this time the solution is to eradicate book production.
This is the part where people usually say, āAI is theft.ā
Look, I get the frustration. AI has moved fast, and the transparency around training data? Not great. Truthfully when I made this deck, I wasn't aware of all of the details about it.
When photography emerged, traditional artists accused photographers of ācheatingā because it captured reality instantly instead of requiring skillful, time-intensive painting.
Some argued it was stealing the soul of a subject, and many feared it would make painting obsolete.
Copyright concerns arose when painters accused photographers of copying their compositions without permission.
Lawsuits were filed against early record companies for using compositions without paying composers, leading to early copyright laws in the music industry.
But does AI have the right to learn from publicly available images the same way human artists do? Thatās where things get murkier. I'm very curious to hear how the court cases will play out, but I think in terms of the way our legal system works, the AI companies will likely pay some amount in damages for specific places they illegally took training data, as they should, and then lots of places where it gets deemed acceptable.
Because if weāre being honestāall art is built on influence. Painters study other painters. Musicians borrow from other musicians. Creativity is, by definition, a remix. The difference is that AI does this at a scale weāve never seen before. Human artists don't instantly absorb millions of works and regurgitate them in any way imagined.
I do genuinely wonder if the scale makes a difference and the extent to which the nuances matter. For example, just like in our current legal system, making a near identical replica is illegal but being inspired is fine. This is up to a conjunction of the ceator and the provider to safeguard, and the law to regulate. Does that inspiration/replica model change just because it's thousands, millions, or billions of images instead of a single one?
Do I think AI companies should compensate artists? Yes, in many cases. If an artistās name is used directly, that crosses the line into influence, and if it's the sole influence (which frankly it never is), that gets closer to copying. I'd like to see a system that uses the percentage of influence as a mechanism of paying out contributors. This may seem like a something to be angry about not existing, but it's worth keeping in mind the scale at which these systems were trained.
If a billion references were included, then a training art piece influenced an image created with the system about one-bilionth (could be less if it's a medium not explicitly called for, or even asked not to be included). The creation of that image might be worth anywhere from $0.01 to $5.00 given most stock photography is only $3-5.00 and is directly created by a creator. It this example, it would then take a billion productions for an artist to get the payout of one unit, of $0.05 to $5. So this is not a silver bullet for artists who had their content used for training. If your work was specifically used by name or image 1000 times in a year, let's say, with a 30% influence, that could might make you $1,500, which is significant.
There is a lot to figure out. I'm really hopeful that it's possible. Much of the issue is that these systems are black boxes, meaning we don't even know what is influencing them or how they're working. We can do our best to make educated guesses though. If you'll entertain this mild tangent; IĀ know blockchain gets a lot of hype and a lot of disillusioned looks, as well, but AIĀ imagery is an area where I think it could provide real value in tracking and regulating AI created content in a distributed fashion, to keep the system honest and know what kinds of devices and software is.
Itās all a real conversation, and I donāt think the answer is simple. I think there are murky territories that the AI companies and the individuals who use them can go into or not. Some of it I was not aware of when I got started, others IĀ was. IĀ do feel good about how IĀ used the technology, but there are valid concerns about the legitimacy of the technology altogether.
People ask, why not just hire an artist to do it by hand?
The short answer? If AI didnāt exist, I would have made the art myself anyway.
The longer answer?
1ļøā£ Iām an artist. This isnāt about outsourcingāI wanted to control my own creative process. I think if it makes people more comfortable, I act in some ways a curator, in other ways an art director, and in other ways as an artist or designer.
2ļøā£ I couldnāt afford it. Hiring another artistāeven at a modest rateāwould have cost more than I could afford. Perhaps I shouldn't have made a deck then. I think it's an idea that makes tarot more accessible, and I think that personally deserved to be created. It's possible this reduces demand for some artists, but I think it creates opportunities for many more people, and I think overall more things will get created and consume rather than less. This may be an issue in terms of environmentalism, but in terms of economics and job loss, I think it means more jobs, more independent creators, and more ideas that can be financially viable and exist in the world.
3ļøā£ And to be clear? At the point of writing this, over many years, I havenāt made money on this deck. I hope to make money. If we count time, Iāve lost a lot of money, as Iāve worked on this full-time for many months and part-time for years.
4ļøā£ The project required endless iteration. I needed to be able to tweak, refine, and reworkāsomething I wouldnāt have been able to do if I was just handing it off or even working with another artist.
So no, I didnāt use AI because it was the āeasyā way. I used it because it was the only way I could afford to create the deck I envisioned.
Here too, this is not new. Before the printing press, books were hand-copied by scribes, often monks. When printing allowed mass production of books, scribes lost their livelihoods. We aren't angry now at people for printing books. Writers and scholars worried about unauthorized copies and plagiarism, a precursor to modern copyright concerns.
AI is disruptive, and people have strong opinions about it. But this isnāt just about artāitās about design.
Design is about function. Itās about communicating an idea effectively. And tarot? Tarot is design. Every image is meant to evoke something, to represent something, to instantly tell a story.
ā
IĀ get some of the outrage. Some of it I didn't know when I started. Others I took into consideration and did the most earnest and honest job I could with it and I feel good about the way I did it and what I created.
It also feels to me that so much of the outrage is the same story that we've been telling for centuries about new technology ruining things, that we then become completely reliant on, and I believe for the better in a not too distant future.
The fact that children and people who have not spent their life perfecting a skill can create exactly what is in their mind in an afternoon after a life of not being able to, is truly a miracle that will lead to so many amazing inventions, stories being told, and worlds being created that died in the minds and hearts of amazing souls since the beginning of time. So when people say things like, "it's soulless", I don't agree. What the person puts in and how they shape it is their intent and their expression. They're breathing magic through their words, concepts, language, and vision instead of their hands and vision alone. There are "good" and "bad" artists, meaning skilled and not-skilled, visionary and not visionary, and there are people who create with AI using creativity and imagination, and those who produce the same tired work.
For me, AI was a toolānot a replacement. A way to create something that feels intuitive, immersive, and true to the Millennial experience. Regardless, it does say on the back of the box it was made in-part using generative AI so like a nutrition label, people know what's inside.
And at the end of the day? You donāt have to agree with me, but if you made it this far, at least now you knowĀ another perspective on the discussion.
