Their employer Anthropic, the company behind a new AI language model Claude, has made history with their settlement. They will have to pay out $1.5 billion in a class action lawsuit, the largest payout in U.S. copyright law history. This settlement comes from the case “Bartz v. Anthropic.” In this lawsuit, the tech giant was caught stealing millions of books from so-called “shadow libraries” to train its AI systems. The deal aims to resolve outstanding legacy claims from up to 500k creators. Each screenwriter will receive a guaranteed minimum payment of $3,000.
The lawsuit against Anthropic harkens to a larger negative shift within the tech industry. Similar legal actions have recently been brought against other behemoths such as Meta, Google, OpenAI and Midjourney. These companies are racing to amass extensive written material for training their large language models (LLMs), which power various AI chat products, including Anthropic’s Claude.
To commemorate the occasion, we have a joint statement from Aparna Sridhar, deputy general counsel at Anthropic, about why this settlement is important.
“Today’s settlement, if approved, will resolve the plaintiffs’ remaining legacy claims,” – Aparna Sridhar.
In June, Judge William Alsup ruled favorably and decisively in Anthropic’s favor. In justifying their actions, he claimed training AI on copyrighted works is legal and fair use. Yet, of everything that made him so upset that it resulted in this lawsuit, he was most touched by the piracy problem. His decision showed an increasing judicial awareness of the ethical consequences of using copyrighted content without authorization.
In particular, Judge Alsup emphasized Anthropic’s purpose in the use of copyrighted works. He said the company aims to create one-of-a-kind outputs, rather than just replicating any content already out there.
“Like any reader aspiring to be a writer, Anthropic’s LLMs trained upon works not to race ahead and replicate or supplant them — but to turn a hard corner and create something different,” – Judge William Alsup.
The lawsuit against Anthropic brings to light the increasing pressure that copyright laws are facing as they intersect with the fast-paced evolution of AI technologies. Companies such as Anthropic are in a race to develop their AI technology. Increasingly, they are perhaps underestimating public sentiment and scrutiny for how they procure training data.
Not only does this case resolution create a strong precedent. Specifically, it will inform how tech companies should deal with copyrighted works when they train their AI models. That result will have an impact on other lawsuits and settlements that follow between the government and other big tobacco defendants.
