For me, it all began with a sign at a Burbank rally last summer. In thick black marker, someone had sketched Bender from Futurama while carrying a sign that said, “Leave Animation to the Humans.” It was a small joke that told you everything about who was going to be there that afternoon, and it was beautifully framed. In the abstract sense, these weren’t tech skeptics. They had spent twenty years learning how to draw, and now they were being told that using their hands was optional.
By all accounts, the Stand With Animation rally was the biggest in the Animation Guild’s history. I can’t stop thinking about that detail. During regular contract negotiations, guilds typically don’t break records. Something has changed, and it feels more like a generational refusal than a labor dispute. The Mitchells vs. the Machines director Mike Rianda has been candid about what he has heard in studio rooms: executives openly considering laying off half of their staff. It remains to be seen if they can truly pull it off. The fact that they are saying it aloud is crucial.

Karla Ortiz and her co-plaintiffs’ lawsuit against Stability AI and Midjourney, meanwhile, overcame a barrier that most observers thought it would not. In other words, the AI companies must now reveal what they used to train their models because the judge allowed it to proceed to discovery. Brian Merchant was informed by Ortiz that she had opened “a nice big bottle of sparkly wine.” It’s a tiny, incredibly human detail. A bottle of something cold on a Friday night, years of labor, and an uncertain future. These fights actually appear like that.
Tech coverage frequently frames this as nostalgia versus progress, which completely misses the mark. According to a 2022 Oxford study, machine learning is a tool, not a replacement, and it is impossible to transfer creative decisions that are based on lived experience into a dataset. A follow-up from Barcelona in 2026 went even further: In a creative imagination task, researchers found that unguided AI performed the worst when compared to visual artists, regular people, and AI image models. Not near. Finally. It’s the kind of discovery that ought to shatter some Silicon Valley confidence, but it most likely won’t.
The way the artists themselves depict the danger is intriguing. They don’t specifically claim that AI is poor at art. They claim it’s acceptable to lower commissions, eliminate junior roles, and allow studios to act as though the finished product is satisfactory. I was struck by Rianda’s statement that executives will make money for two years before the bottom falls out, at which point the jobs will be gone. That has a subtle, tired realism to it. Don’t panic. Not Luddism in the sense of a cartoon. Just knowing how these cycles come to an end.
As you pass the picket lines, the signs, and the lawsuits, it’s difficult to ignore the fact that those closest to the work are the ones who are least impressed by the machines. That most likely has some significance. The question of whether the courts, studios, and venture capitalists will listen to it is still up for debate.
