The lawsuit filed by USA Today Co., Inc. on October 8, 2026 claims its papers make up over 160,000 entries in OpenAI's WebText dataset and over 122 million tokens in a 2019 Common Crawl snapshot