Artists Take Legal Action Against AI Misuse, Achieving Notable Victories
When a dataset revealing the works used to train AI was made public, Kirk Wallace Johnson, like many artists, searched for his name out of curiosity. To his dismay, he found it. His nonfiction books, including *The Feather Thief* and *The Fishermen and the Dragon*, which he dedicated five to six years to researching and writing, had been pirated and utilized to train a chatbot. This discovery stirred a mix of emotions in him: anger at the blatant theft, concern for the implications for writers, and a strong desire for retribution against the corporations profiting from his intellectual property.
Johnson took action by contacting a law firm already pursuing a case against Anthropic on behalf of authors, viewing their lawsuit as a stand for all creators. He is among a growing number of authors, musicians, illustrators, and various artists who are challenging AI in court. Their lawsuits primarily focus on copyright infringement, though some also address violations of terms of service. While some cases have dragged on for years, others have reached settlements more swiftly. Throughout this process, artists have experienced both victories and setbacks, particularly regarding the interpretation of fair use.
Illustrator and cartoonist Sarah Andersen was one of the first to vocally confront the AI giants. She describes her webcomic *Sarah’s Scribbles* as a deeply personal project, a culmination of her education and life experiences. In a 2022 editorial, she expressed how her work is a reflection of her life. Alongside fellow artists like Karla Ortiz and Kelly McKernan, she filed a class action lawsuit against several AI companies, which has been slowly progressing through the courts since January 2023, shortly after the launch of Stability’s image generator, Stable Diffusion, and Midjourney. What began as a curiosity in generative AI has escalated into a significant issue.
Other artists, inspired by Andersen’s actions, have initiated their own legal battles against major AI players, including Meta, Google, Anthropic, and the AI music generator Suno. Many artists remain hopeful about the outcomes of their cases, believing their efforts could help establish legal boundaries. However, they also express serious concerns about the mindset of AI companies. Johnson notes, “This does not seem to be a bus driven by a bunch of sane sober thinking people, and we’re all stuck in it.”
A common sentiment among artists is that the leaders of Big Tech and those developing AI models lack a fundamental understanding and respect for art. Andersen has voiced feelings of violation, stating that it reduces her life’s work to mere algorithms. Author Andrea Bartz, known for her novels like *We Were Never Here*, echoed this sentiment, expressing shock and alarm at seeing her years of hard work treated as just another data point for AI training.
Sam Kogon, who wishes to be recognized for his music rather than his legal battles, is the lead plaintiff in a lawsuit against Google’s Lyria AI music engine. He argues that AI devalues artistic work by making it freely available, which could disenfranchise many musicians. Kogon also criticizes AI-generated music as fundamentally dehumanizing, stating that creating artificial music is “the most anti-human thing you could do.”
Kogon and other independent musicians are accusing Google of breaching its own terms of service, which sets their case apart from many others focused on copyright infringement. Their argument centers on Google’s alleged misuse of its Content ID system and YouTube data to train Lyria and ProducerAI. While Google has largely refrained from commenting on the specifics, it has filed a motion to dismiss, claiming broad rights under YouTube’s terms of service to reproduce and create derivative works.
Kogon describes this situation as a “pure bait and switch,” highlighting that YouTube frequently alters its terms of service, which users must accept without negotiation. He argues that Google’s stance implies that any content uploaded can be used for training purposes, regardless of the original context. Krystle Delgado, an entertainment and IP lawyer, challenges Google’s claims, noting that most users do not expect to relinquish rights to their content when using YouTube.
Artists I spoke with view this as an exploitation of Google’s power, as opting out of such a massive platform is not feasible. Independent artists and the creative workforce are particularly vulnerable. When novelist Richard Kadrey and others sued Meta for using their works to train its Llama AI without consent, their complaint highlighted the potential harm to emerging authors. They argued that while AI-generated books might not impact established authors, they could hinder new voices from gaining recognition.
Johnson emphasizes that focusing solely on famous authors misses the broader issue. He believes AI may not replace great art but could produce mediocre works that fill the market, affecting many writers and screenwriters. The judge in Kadrey v. Meta dismissed several initial claims due to a lack of evidence for market harm, but a narrower set of claims regarding copyright infringement is still progressing through the courts.
In Bartz v. Anthropic, the court found that the company violated copyright laws by using pirated ebooks for training its AI. The settlement, amounting to $1.5 billion, marked a significant moment, as it was the largest ever in a copyright case, and the company agreed to destroy its collection of pirated materials. However, complications arose regarding the millions of secondhand books acquired for training under Project Panama, which the judge ruled could be used under fair use due to their transformative nature.
Bartz expressed her disagreement with the ruling, hoping future courts will recognize the implications of such decisions. She pointed out that even libraries cannot scan and lend out physical books as ebooks. Nonetheless, she acknowledges the importance of her case as a landmark moment for holding AI companies accountable for using artists’ work without permission, viewing the settlement as a significant victory for creatives.
Delgado, who is spearheading the case against Suno and Udio for independent musicians, shares a similar optimism. Despite setbacks in the cases against Anthropic and Meta, she believes the tide is turning in favor of artists. “Right now with these companies, they’re really nervous,” she notes, observing that both the courts and public opinion seem to be shifting in their favor. Polls indicate a growing demand for transparency in AI practices.
Even with potential legal victories, creators and lawyers express concerns about the sustainability of artistic careers amid the rising influence of AI. Bartz warns against the pervasive narrative that AI is an inevitable force, urging people to consider the damage these companies inflict on the arts, critical thinking, and the economy as they continue to accumulate power and wealth.
