AI company Anthropic is set to pay $1.5 billion to thousands of authors in a significant copyright infringement settlement. However, this resolution brings mixed emotions for some authors involved.
Mixed Reactions Among Authors
Bestselling author Charles Graeber expressed both pride and reservation about his role in the lawsuit. Despite being part of a significant collective action against a tech giant like Anthropic, Graeber hesitates to call the outcome a total success. He expects compensation for two of his works used by Anthropic’s AI model, including ‘The Good Nurse.’ Yet, he reflects on the personal cost in terms of time and missed opportunities.
Settlement Details
The San Francisco federal court approved the $1.5 billion settlement, initiated by a class action lawsuit against Anthropic. The lawsuit stemmed from allegations that the AI company used digitized books to train its chatbot, Claude, without compensating the authors. Despite the settlement, the distribution of funds means authors receive only a portion of the total, shared also with publishers and lawyers.
Implications for AI and Publishing
Some see the lawsuit’s outcome as empowering for AI companies, particularly under the fair use doctrine. Aparna Sridhar from Anthropic stated that training AI on books is fair under copyright law, provided authors are compensated. This viewpoint was supported in a similar case involving Meta, where they successfully argued the fair use defense.
Challenges from International AI Markets
The global stage presents additional challenges. Attorney Dale Nelson highlights how AI companies abroad, especially in China, may not adhere to the same legal standards, using different methods to train AI models without direct book copying. Andrea Bartz suggests international agreements as a potential solution to these challenges.
Licensing as a Future Path
Licensing appears as a promising approach to resolve future disputes. Umair Kazi from Authors Guild advocates for licensing as a way to legally facilitate AI training. Such agreements could enable authors to control how their work appears in AI outputs, potentially limiting the creation of derivative works. Current agreements exist, such as Perplexity AI’s deal with media companies, but widespread adoption remains uncertain due to ongoing legal battles.
Leave a Reply