When The Atlantic released a searchable list of works used to train large language models, author Kirk Wallace Johnson traced his own titles—*The Feather Thief* and *The Fishermen and the Dragon*—to the dataset. He described the discovery as a "cocktail" of anger and betrayal and promptly engaged Susman Godfrey, the firm already representing authors in a suit against Anthropic. Johnson’s case frames the broader movement of creators who feel their intellectual property has been siphoned by AI firms for free.
Illustrator Sarah Andersen, best known for the webcomic *Sarah’s Scribbles*, joined a class‑action suit that targets Stability AI, Midjourney, DeviantArt and Runway AI. Filed in January 2023, the complaint argues that the companies violated copyright by training image generators on millions of unlicensed artworks, including Andersen’s. Her public statements characterize the infringement as a reduction of a lifetime’s work to a set of algorithmic patterns.
Musician Sam Kogon leads a separate lawsuit against Google, accusing the tech giant of breaching its own YouTube Terms of Service when it harvested audio uploads to develop the Lyria and Producer AI music engines. Kogon’s legal team contends Google used the Content ID system and YouTube data without consent, effectively giving away musicians’ creations for free. Google counters that its terms grant broad rights to reproduce and create derivative works from uploaded content.
Author Andrea Bartz, whose thrillers *We Were Never Here* and *The Spare Room* are among the works alleged to have been used by Anthropic, also filed suit through Susman Godfrey. The case resulted in a $1.5 billion settlement—the largest ever in a copyright dispute—along with a court order for Anthropic to destroy its pirated ebook collection. However, a judge later ruled that Anthropic’s use of legally purchased books, part of a project dubbed “Project Panama,” qualified as fair use because the training process was “quintessentially transformative.”
Other creators, including novelist Richard Kadrey, comedian Sarah Silverman and writer Christopher Golden, have sued Meta for incorporating their books into the Llama model without permission. While a federal judge dismissed many of their market‑harm claims, the remaining copyright‑infringement allegations continue to move through the courts.
Legal analysts note that the outcomes of these cases could reshape how AI developers source training data. So far, judges have shown a willingness to scrutinize the scope of fair use in the context of machine learning, and public opinion appears to favor the artists. Yet many creators remain uneasy, fearing that even if they win individual battles, the sheer volume of AI‑generated content could erode their ability to earn a living.
Cet article a été rédigé avec l'assistance de l'IA.
News Factory APP - actualités agentiques pour booster votre SEO et AEO.