AI
Artists are taking AI companies to court—and landing some significant victories
A wave of lawsuits from creators over unauthorized AI training is reshaping how courts view copyright infringement, though the broader battle for artists' livelihoods remains uncertain.
The Verge
When The Atlantic released a searchable database revealing works used to train AI systems, author Kirk Wallace Johnson searched for his name and discovered his own books—including The Feather Thief—had been pirated and fed to chatbots without permission. His reaction crystallized a growing movement: writers, musicians, illustrators, and visual artists across disciplines are turning to the courts to challenge the tech giants behind generative AI.
The legal landscape takes shape#
Johnson's discovery spurred him to join Susman Godfrey's lawsuit against Anthropic, viewing the case as a statement against what he calls an insane, uncontrolled industry. He is far from alone. Dozens of creators have filed suits targeting companies like Google, Meta, Anthropic, Midjourney, and Stability AI, primarily on copyright grounds, though some pursue alternative legal theories like terms-of-service violations. The cases span different industries—visual artist Sarah Andersen's 2023 class action against AI image generators has been advancing through courts since January 2023, while musicians like Sam Kogon are challenging Google over its music AI tools using terms-of-service angles.
These suits reflect a shift in creative industries' attitudes. When Stable Diffusion and Midjourney first launched, generative AI was a niche curiosity; now it has become a policy priority, with lawmakers debating AI regulation as a national security concern.
When artists win: The Anthropic precedent#
The most significant victory came in Bartz v. Anthropic, where the court found the company had violated copyright law by using pirated ebooks to train Claude. Anthropic settled for $1.5 billion—the largest copyright settlement ever—and agreed to destroy its collection of illegally obtained books. This marked "the first time a large AI company has been held accountable," according to Andrea Bartz, the lead plaintiff and novelist, who described the outcome as a potential "first major win for creatives."
However, the ruling's fine print matters. Judge William Alsup determined that Anthropic's use of legally purchased books from its "Project Panama" initiative qualified as fair use because it was "quintessentially transformative." Bartz strongly disagrees with that portion, arguing the logic would never apply to libraries scanning physical copies. Still, the settlement sends a message that copyright protections are not dead.
