In January 2023, Joseph Saveri and I filed the initial Andersen v. Stability complaint on behalf of three visual artists—Sarah Andersen, Kelly McKernan, and Karla Ortiz. The complaint challenged the legality of certain AI image generators, including those made by Stability AI, Midjourney, and DeviantArt. In October 2023, Judge William Orrick dismissed most of our claims. But in November 2023, we filed an amended complaint (adding seven more fantastic artists as plaintiffs, and Runway AI as a defendant). Last week, Judge Orrick issued an order allowing our core claims to proceed, including all our copyright-infringement claims.
No, this doesn’t mean we win. On the contrary, it only means we get to start: we can begin discovery and put the case on a path to trial. But it’s an important step forward for the case, which had otherwise been paused for over 18 months.
Some coverage of the decision in the Hollywood Reporter. Also some encouraging words from Brian Merchant—author of Blood in the Machine, an excellent history of the Luddite movement—who says that “many predicted the lawsuit wouldn’t get past this crucial hurdle” and that our success in doing so is “a beautiful thing.” Merchant also notes:
Neither the legal case nor the organizing [of the Animation Guild, a labor union] are campaigns against the technology itself, after all, but efforts to wrest consent over how the technology is used in their lives.
While you’re waiting for Merchant’s book to arrive, some major points of the historical Luddite movement—as distinct from the latter-day tech-bro smear—are covered by Tom Humberstone’s web comic.
I’m honored to be representing over 20 creators in seven lawsuits challenging generative AI. Still, the Andersen case has always had special weight for me. For one thing, the case is probably the most technically knotty of the seven. It raises unique factual & legal issues—about which another time.
But also—my thinking about these cases initially arose from noticing that my own work was in the training datasets for all these commercial AI models: my writing, my software code, my visual images. Still, I identify most strongly as a visual artist and designer. That’s the work that’s motivated (and paid for) everything else I’ve done—including law school. Having carved out this niche over many years, it was all too easy to foresee how generative AI based on illegal practices could knock off or permanently dilute that work.
A visual-artist friend of mine recently told me she found it bewildering that AI proponents frequently depict visual artists as “gatekeepers”—the theory, I guess, is that visual artists’ objections to generative AI arise from some antidemocratic despair that others might be able to use AI tools to artistically express themselves. As my friend put it, “They don’t understand—when I’m making art, that’s basically the only time I feel normal.”
I understood. Generative AI represents not just an economic threat to artists, but also a deeper, personalized encroachment on a certain intellectual & emotional habitat. I’ve been self-employed for my entire professional career. Not because I have some antipathy to being employed by others. Rather, because the habits that make my work distinctive and profitable as an independent designer are not well tolerated in the workplace. For instance—I don’t like asking for permission. I like to make things and put them in the world and see what happens.
So let’s see what happens.