Good morning. I’ve been struggling to get back to work all this week after being sucked into some new novels and briefly forgetting that the rest of the world exists. Are book hangovers a thing? If so I definitely have one.

It was a busy week in the art world and in my studio though, so let’s get into it.

Well, I can’t tell you that I didn’t see this coming.

Following her op-ed in the New York Times, Sarah Andersen and several other artists have filed a lawsuit against the companies behind three major AI art generators: Midjourney, Stable Diffusion, and DeviantArt’s new tool: DreamUp.

Their chief complaint? That the models used to train these tools (see this newsletter edition for more background) have used their artwork without consent or compensation. In the lawsuit, Andersen and the other plaintiffs cite a study that found that 47 percent of the training dataset is made up of images from Shutterstock and Getty, as well as user-generated content from Pinterest, Flickr, Tumblr, and DeviantArt. An estimated 1 in 50 images, totaling over 3 million works, were taken from DeviantArt.

Additionally, the lawsuit makes a claim that Andersen brought up in her op-ed: these tools have allowed art to be generated in the style of individual creators. As such, they’re seeking financial compensation for potential lost revenue.

Getty, a popular stock photo site with more than 15,000 photos in the training datasets, has brought a separate lawsuit against Stable Diffusion’s parent company for copyright violations. In a recent article in The Verge, Getty’s CEO Craig Peters says that Stable Diffusion “made no outreach to Getty Images to utilize our material, so we’re taking an action to protect our and our contributor’s intellectual property rights.”

“These generative models need to address the intellectual property of others, and that’s the crux of it,” Peters continued, “and we’re taking this action to get clarity.”

The main defense the AI firms are relying on is the fair use doctrine. The fair use doctrine permits limited use of copyrighted material without having to seek permission from the copyright holder. It’s what allows streamers on Twitch to cover news stories, and your favorite commentary channels on Youtube to show copyrighted material like clips from movies, TV shows, etc without being taken off the platform.

Without the fair use doctrine, the internet as we know it wouldn’t exist: there would be no intricately layered in-jokes, memes, insightful commentary, fan-cams, and mashups. See this Tom Scott video for more in-depth coverage.

The fair use doctrine is absolutely a good thing, but there are limitations.

One such limitation to the fair use doctrine is the extent to which the use of the copyrighted material is transformative. Are you adding something new or changing it in a meaningful way?

Uploading a pirated version of Legally Blonde on Youtube is obviously not okay, but if you made a 20 minute video essay analyzing the ways in which Elle Woods is the best version of a #Girlboss, using clips from the movie along the way as evidence for your argument? Well, that’s fine. You’re adding something new, the use is transformative.

I’m sure you can see where we’re going with this. AI art models combine so many individual images together that it’s often (but not always) extremely transformative. (My take? going this route will not bear fruit for the artistic community.)

That brings me to another factor in fair use cases: the effect upon the original work’s value.

In the not so far off future where someone can easily replicate an individual artist’s style, it’s obvious that an artist’s ability to monetize their own work is severely adversely affected.

AI would essentially present a direct substitution to the original artwork, likely sold at a far lower price than the original. After all, they could churn out new work at a breakneck pace, not bound by the constraint of the original artist’s creative process.

The current question is whether the copyright owners, people like Sara Andersen, can prove that their income has been affected.

If they can, then personally I’m optimistic about the outcome of the case. But if they can’t, their failure could result in terrible legal precedent for artists.

I find it unlikely that either side will get the decisive victory they’re hoping for, at least in the long term. AI art isn’t going to just magically go away. In fact, I’m far more worried about the development of AI in a more general sense, not even as it applies to the art world.

My realistic expectation is that companies like LAION and models like MidJourney and Stable Diffusion will have to do the obvious thing and prohibit the wholesale copying of the style of living artists like Sara Andersen.

I wouldn’t even be surprised if Getty manages to win an outcome such that the work of copyright holders like Andersen and Getty have to be purged from the training datasets, and more regulations are put in place for image scraping going forward.

This week on the main channel I posted a video going into detail on my plan to hopefully generate $150k in revenue for my business this year. It’s a pretty ambitious plan, but hopefully doable.

(If I ever want to stop setting my money on fire by paying rent and actually become a homeowner in this current market, focusing on increasing revenue is essential)

But some viewers of that video expressed some doubts on my business model because, as they rightfully noted, not even the majority of revenue in this plan would result from direct art sales.

And they’re right, it doesn't. In the video, I explicitly talk about how I’m hoping to make ~$30k of revenue from my online shop. Last I checked, that’s not $150k.

I’ve said this before, and I truly do believe it so I guess it’s worth repeating: I think placing all of your eggs in one basket is a mistake. I think solely relying on direct art sales for your revenue is a mistake.

You don’t need to be a content creator like I am, it’s not for everyone! I’ll be the first person to acknowledge the downsides. But pursuing alternate means of making it work as a creative isn’t a bad thing.

Personally, I really like how I can be extremely experimental with my work right now as a result of not having to rely on direct sales. It means I can follow the dopamine and make art that I’m truly passionate about.

Obviously I want to increase the revenue that’s directly stemming from sales of my artwork. But I’d argue that revenue generated by me sharing my art and creative process online, without people even having to do anything other than watch the process, is still a more than acceptable means of making a living as an artist.

Artists in the modern era have to wear a lot of hats: we’re business strategists, marketers, content creators, copywriters, and so much more on top of being artists. Why shouldn’t you aim to make money from your marketing efforts on platforms like Youtube? If I have skills and wisdom to share with the artistic community that consumers find valuable, why shouldn’t I use that to generate revenue?

I find critiques of this nature to be unsatisfactory. If you don’t want to follow my advice or emulate my own journey, you shouldn’t! You should follow whatever path works best for you and stay true to yourself.

And that’s it for this week!