After losing on its motion to dismiss, Meta filed its answer to Kadrey’s Third Amended Consolidated Complaint.
One important admission is that the plaintiffs’ books were in the Books3 dataset. Meta also admits: “it used data from LibGen, Z-Lib, Sci-Hub, Common Crawl, and data collected using its web crawler Spidermate to train one or more of its Llama models.”
46. The allegations in this paragraph state a legal conclusion to which no response is required. To the extent a response is deemed required, Meta denies that it infringed Plaintiffs’ alleged copyrights. Meta admits that substantially all of the text from some of Plaintiffs’ books appear in the Books3 dataset; that it used data from LibGen, Z-Lib, Sci-Hub, Common Crawl, and data collected using its web crawler Spidermate to train one or more of its Llama models; that it put certain URLs on a blocklist to not be scraped; and that some of its employees expressed their opinions regarding the use of data sourced from certain websites to train its LLMs. As to any remaining allegations in paragraph 46 regarding Meta, except as expressly admitted, Meta denies those allegations. Meta declines to adopt the definition of “Infringed Works,” which calls for a legal conclusion. Meta lacks knowledge and information sufficient to form a belief as to the truth of the remaining allegations set forth in paragraph 46, and on that basis denies the same.
Related Stories
