Greetings from the Digital Frontier: A.I. Art in Card Design

Since we’re technically still in the month of January, I’ll take this opportunity to wish you all a happy and healthy 2023! This blog has been on a long hiatus due to time constraints — in part from having moved a year ago — but I hope to resume posting here at least intermittently.

This past holiday season, I decided to take a different approach to sending out my holiday cards. While I still placed orders with some of my favorite card designers, there were a couple of visual concepts I had in mind that I wasn’t easily finding in the existing offerings I came across, so I added to the mix my first-ever custom-created designs in the Temporal Treasures line. I was envisioning one image of some wolves dancing with glee on a snowy moonlit night and another of a baby harp seal lying serenely on an ice floe beneath an aurora borealis…so that’s what I made (see composite header image above). But in the spirit of full disclosure, I had quite a bit of help with bringing these images to fruition…I could not have done it alone.

Enter OpenAI’s Dall-E 2, one of several artificial intelligence (A.I.) art generation tools that caused an uproar among flesh-and-blood artists this summer when an A.I.-generated work won an art contest. These tools’ artistic capabilities are developed by aggregating large numbers of existing artworks available online, typically and unfortunately neglecting to seek the artists’ consent beforehand, and associating keywords with them as the basis for iterative training to partially emulate — but not duplicate — those works. By entering strings of keywords into a language prompt, without so much as lifting a paintbrush or pencil, users of these tools can generate their own pieces that to some degree are distinct from any others in the world. Users are legally allowed to take credit for the pieces they generate and do with them what they wish, including selling them.

Objections from the artist community have been strong and swift — and, I might add, very understandable: If a tool’s capabilities are based on the creations of others without their knowledge and allow users to almost instantaneously build on them, isn’t that stealing? If artists took hours or years to create their pieces and an A.I. user swoops in to generate a dazzling tableau in a matter of seconds, isn’t that cheating? If a piece of art is generated by typing in strings of words rather than being constructed by a human artist, is it legitimately art at all? A.I.-generated art in its current form has entered murky ethical territory and unleashed a tidal wave of intellectual property dilemmas. In fact, a group of artists has just filed a potential class-action lawsuit against three A.I. art generation makers that could have major implications for how content is produced by A.I. tools in the future.

So in light of all that, why would I still choose to use Dall-E 2 for my cards?

While I recognize its problematic aspects, I also see incredible potential to democratize the creation of visual art and make its production more accessible to those who, for any number of reasons, would otherwise be unable to pursue it and may not necessarily seek to profit from it. Although I introduced two new card designs using art I made with Dall-E 2, I didn’t sell them — the cards in question were shared only among friends — and I was very upfront about not having created the images from scratch. (I also used Photoshop to overlay the lettering on top of the images.)

But the train of profiting from A.I. art has already left the station. Even if I had sold cards bearing my A.I. images (which I haven’t ruled out doing in the future), there would have been no prohibition against that in the eyes of the law as it stands today. A.I. art does not violate current copyright restrictions because each piece that is generated contains unique characteristics. Had I found exactly the kinds of images I was envisioning on the open market, I would have gladly paid for them — but the fact that the A.I.-generated art filled a specific gap for me gives further credence to the idea that it contains at least a spark of originality…goodness knows I’ve seen plenty of A.I. art that is unlike anything else I’ve previously encountered. (Also, as it happens, Dall-E 2 is not currently among the A.I. art generation makers named in the lawsuit.)

I would also argue that those of us who lack the talent, training, or time to make visual art using more conventional methods have the opportunity to develop another kind of skill: what has become known as “prompt craft” — the ability to cobble together the right combination of keywords to produce a favorable result, which could be viewed as an art form itself. It is also interesting to ponder the possibilities for people who have vision-related disabilities that may limit their participation in the direct production of detailed visual art via standard means, but who now have this new creative door open to them.

The next few years will be formative ones for the world of A.I. art as the laws and ethics around it become more clearly defined. In the meantime, I’ll be following the developments with great interest on what is sure to be a wild ride.

Published by clairesterling

Please see my personal website at clairesterling.com for information about me.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

A Life Of Stationery

Write it, Draw it, Live it...

Bournemouth Girl

Planning the life you want

The Paper Nerd

Veni Vidi Scripsi

%d bloggers like this: