Rep. Jamie Raskin (D-MD) said that AI company Anthropic used three of his books [1] without permission to train its Claude model.
This confrontation highlights the escalating legal and ethical tension between generative AI developers and content creators. As AI companies scrape vast amounts of data to improve their models, authors and artists are demanding compensation and legal protections against what they describe as intellectual property theft.
During a congressional hearing at the U.S. Capitol, Raskin delivered an opening statement focusing on the lack of transparency in AI training sets. He said that AI firms are profiting from the work of creators without providing any payment or seeking consent. The congressman specifically named Anthropic, saying that the company utilized three [1] of his published works to develop the Claude AI system.
Raskin called for the immediate drafting of new copyright-related legislation to address these gaps in the law. He said that current regulations are insufficient to protect authors, artists, and consumers from the unchecked expansion of AI training practices. The representative said that the current model of "scraping" data constitutes a violation of creator rights.
Legislators are currently weighing how to balance the rapid growth of the tech sector with the preservation of intellectual property. Raskin said that the industry must move toward a system of transparency and fair compensation. He urged his colleagues to prioritize laws that would require AI companies to disclose their training sources and pay royalties to the original creators.
Anthropic has not yet provided a detailed public response to the specific allegation regarding Raskin's books. The hearing continues to explore how the U.S. government can regulate the intersection of machine learning and copyright law to ensure a fair marketplace for human creators.
“AI companies are profiting from creators' work without compensation”
This incident underscores a critical legal pivot point for the AI industry. If high-profile lawmakers successfully push for mandatory disclosure of training sets and royalty payments, it could dismantle the 'fair use' defense currently employed by many AI labs. Such a shift would likely increase the operational costs of training large language models and create a new licensing economy for digital content.
