• An AI copyright infringement lawsuit against Anthropic has been settled with U.S. authors.
• The lawsuit alleged unauthorized use of copyrighted works to train the AI assistant, Claude.
• The terms of the settlement remain undisclosed, but commitments to improve transparency were made.

Artificial intelligence company Anthropic has settled a class action lawsuit with U.S. authors who claimed the company infringed their copyrights by training its AI models using their works without permission. This lawsuit, which was initially filed in 2024, accused Anthropic of illegally downloading thousands of books—including fiction, non-fiction, and academic texts—to train its AI assistant Claude, potentially exposing it to billions of dollars in damages, according to Reuters and LA Times.

In June, a U.S. District Judge found that while Anthropic's method of using copyrighted material could be deemed "fair use," the company violated copyright law by saving millions of pirated books to a centralized location. The settlement, whose financial terms have not been disclosed, marks a significant moment in the ongoing battle between AI developers and creative professionals, as well as establishing new standards for AI training practices going forward, according to India Times.

Details within the settlement suggest commitments from Anthropic to enhance transparency regarding data sources used for training AI models, including improved opt-out mechanisms for authors who do not wish to have their works used in future trainings. Although Anthropic did not admit any wrongdoing, the settlement indicates its dedication to ethical AI development and respect for intellectual property rights, as emphasized by their attorney Justin Nelson, according to Reuters, LA Times, and India Times.