Anthropic Proposes $1.5 Billion Settlement for Authors' Copyright Claims

The artificial intelligence company Anthropic has put forward a proposal to compensate authors with a total of $1.5 billion after facing legal action regarding the unauthorized downloading of their works. This move comes in response to allegations that approximately 500,000 books and other texts were utilized without permission in the training of their AI chatbot, Claude, through two illicit online databases.

Under the proposed settlement, each affected author would receive about $3,000 (approximately EUR2,500) for their respective works. While the authors involved in the lawsuit have accepted the offer, the agreement is contingent upon approval from a judge in San Francisco. Anthropic aims to avert a trial that could result in significantly higher financial penalties.

Claude is recognized as one of the primary competitors to OpenAI's widely-used chatbot, ChatGPT. The current legal landscape is fraught with multiple copyright claims against various AI companies regarding the use of protected works for training purposes. These AI models require vast amounts of data to deliver meaningful responses to user inquiries.

In the San Francisco case, the presiding judge indicated that Anthropic's use of copyrighted texts might fall under the "fair use" doctrine, which allows for some utilization of protected material if it results in the creation of something new. However, the judge also ruled that the downloading of the two piracy databases did not meet this criterion. Furthermore, it was concluded that Anthropic was aware of the illegal nature of these databases, which could lead to penalties of up to $150,000 per infringed work if the case proceeded to trial. This potential financial risk prompted Anthropic to seek a settlement.

The outcome of this case could set a significant precedent for how AI companies handle copyrighted material in the future, especially as the industry continues to navigate the complexities of intellectual property rights in the digital age.