Anthropic reaches a settlement over authors' class-action piracy lawsuit (www.engadget.com)

Anthropic has reached a confidential settlement in a class-action lawsuit filed by authors over the company’s use of copyrighted material to train its large language models (LLMs). The settlement allows Anthropic to avoid potentially massive statutory damages, as the court found their acquisition of copyrighted works—estimated at around 7 million pieces—to be unauthorized piracy, with fines starting at $750 per infringed work. While the judge ruled that training AI on copyrighted content constituted fair use, the unlawful sourcing of that data opened Anthropic to substantial legal risk. This settlement marks a significant moment in the ongoing legal struggles defining AI training data practices, as few clear precedents have emerged on copyright infringement related to AI models. Coming after prior litigation from the music industry against Anthropic, this case highlights the intense pressure on AI developers to balance innovation with respecting intellectual property rights. The final terms remain undisclosed, but the outcome could influence how companies approach data acquisition and fair use defenses moving forward, impacting both model training methods and the broader dialogue on AI ethics and legality.