Since January 2023, a wave of class-action lawsuits against artificial intelligence giants has been redrawing the battlefield between human creators and generative machines. Illustrators, authors, musicians, and photographers are turning to the courts to fight what they see as the unauthorized use of their work in training models like ChatGPT, Midjourney, Stable Diffusion, and DALL-E.
Cartoonist Sarah Andersen, creator of the popular webcomic "Sarah's Scribbles," was one of the first to take on the AI companies directly, becoming a co-lead plaintiff in a class action against Stability AI, Midjourney, DeviantArt, and Runway AI. Andersen argues that her works were scraped from the web and used to train models without any compensation or consent.
Kirk Wallace Johnson, author of "The Feather Thief," discovered his books in a searchable dataset published by The Atlantic containing works used to train AI systems. "I spent five to six years researching, writing, and investigating" each book, Johnson told The Verge. He described feeling a "cocktail" of emotions: "anger over the brazenness of the theft, worry over what this means for writers, and a healthy thirst for revenge on these massive corporations that have become galactically wealthy" using his intellectual property.
Johnson proactively reached out to Susman Godfrey, the law firm already leading the case against Anthropic on behalf of authors, because he saw their suit as a "middle finger on behalf of everyone that has tried to create something."
The class actions target not just Stability AI and Midjourney but also OpenAI, Google, Meta, and Anthropic. Results so far have been mixed. Some artists have secured significant courtroom wins, while others have lost on fair use grounds — the central defense of technology companies.
In parallel, major record labels are pushing for a global ban on AI-generated songs — dubbed "AI slop" — from music charts, fearing the market will be flooded with synthetic tracks that devalue the work of human artists.
Wired recently reported on a curious phenomenon: baby boomers are gifting their grandchildren AI-generated children's books, with personalized stories based on real photos of the kids — much to the dismay of parents, who find the books poorly written and "morally reprehensible" for feeding children's likenesses into "slop machines."
The landscape is one of a multi-front war. As cases slowly work their way through the courts, the creative ecosystem faces an existential question: how to protect human originality in an era where anything can be replicated in seconds by a machine that learned from the work of millions of living artists?
Broader legal context
This legal battle represents only the opening salvo in what promises to be a prolonged conflict between content creators and AI companies. The core question — whether training AI models on copyrighted works constitutes fair use or infringement — remains unresolved, and different courts have reached different conclusions. Some judges have dismissed claims on fair use grounds, while others have allowed class actions to proceed. This inconsistency creates uncertainty for both sides and makes it difficult for businesses to plan their AI strategies.
What is clear is that the window for voluntary agreements may be closing. AI companies that could have negotiated licensing deals years ago, when lawsuits were first filed, now face the prospect of court-ordered damages that could multiply into billions of dollars. The outcome of these cases will shape not just the economics of AI training data, but the very definition of authorship in the age of generative models — a debate that society has barely begun to confront.
