Anthropic scored a legal win when a federal judge ruled that the training of its chatbot Claude using hundreds of legally acquired books is protected under fair use. The judge, William Alsup, emphasized that the chatbot generates new content rather than mimicking original texts. However, Anthropic is still accountable for using millions of illegally obtained books from the internet. The ruling stems from a class-action lawsuit by authors accusing the company of theft of their works. Despite some progress, the company must still address claims regarding the unauthorized materials used in its training process.
Judge William Alsup noted that, although Anthropic legally purchased books for training its models, it is still liable for the use of illegally obtained works.
Anthropic’s internal documents revealed researchers were concerned about the legality of using pirated texts, leading to a decision to purchase books legally, which doesn't erase prior violations.
Collection
[
|
...
]