AI firms OpenAI, Microsoft and Anthropic face U.S. copyright lawsuits over training data

A coalition of 35 small‑publisher owners representing about 400 community newspapers filed a lawsuit on June 24 alleging that OpenAI and Microsoft scraped and copied their news stories without permission to train ChatGPT and Microsoft Copilot. The plaintiffs argue that the AI products have generated billions of dollars in market value while the newspapers received no compensation. The case joins earlier actions by The New York Times and Alden Global Capital and could set a precedent for compensating the roughly 5,400 local papers that provide most U.S. community journalism.

Separately, more than 100 authors have launched a $75 million lawsuit against Anthropic, accusing the startup of pulling millions of copyrighted books from illegal shadow‑library sites such as LibGen to train its Claude chatbot. The authors cite a prior $1.5 billion settlement against the company and contend that using pirated copies for AI training is outright copyright infringement. Both lawsuits reflect mounting legal pressure on AI developers over the use of copyrighted material in model training.