Anthropic, an AI company, has agreed to pay $1.5 billion to hundreds of thousands of authors as part of a landmark copyright infringement settlement. Some authors have mixed reactions to the outcome. Charles Graeber, a New York Times bestselling author, expressed pride in being one of the main plaintiffs. He noted that a group of authors and publishers could stand against a major entity like Anthropic.
Graeber is among over 300,000 writers who will receive compensation. A federal judge in San Francisco approved the settlement in July. The case stemmed from authors’ claims that Anthropic used digitized books to train its AI model, Claude, without consent or compensation. Graeber expects to receive about $3,100 per book for his works, including ‘The Good Nurse’. He appreciated the affirmation that piracy is a crime but felt conflicted about the lengthy legal battle.
‘I’m much poorer for this settlement, ironically,’ Graeber stated, reflecting on the personal costs of the process, including missed job opportunities.
While the settlement is substantial, it must be shared among many authors and publishers. Over $100 million also went to lawyers. The case still poses questions about the AI industry’s potential benefit. The lawsuit did not deem it illegal for Anthropic to train AI algorithms using copyrighted works if they pay for them.
Aparna Sridhar, deputy general counsel for Anthropic, stated, ‘Training AI on books is fair use under copyright law.’ Most authors and publishers involved accepted their share of the settlement. The company did not address further implications of the ruling.
The issue of AI companies using copyrighted works also involves international challenges. Dale Nelson, an intellectual property lawyer, noted the difficulties of enforcing copyright laws abroad, particularly with companies in countries like China. Chinese AI companies often use ‘AI distillation’, utilizing outputs from U.S.-based models for training.
Author Andrea Bartz expressed concern over the fair use argument, stating it allows AI models to compete with human-created work. She advocates for international AI regulations to protect authors.
Authors Guild director Umair Kazi suggests licensing as a solution. Licensing ensures legal training and allows rights holders control over AI outputs. Some such agreements exist, like Perplexity AI’s content licensing with media outlets.
Kazi noted licensing deals remain rare due to ongoing litigation but sees potential for future adoption.