A class action accuses Adobe of secretly pulling hundreds of thousands of copyrighted books—including titles scraped from pirate “shadow libraries”—to train its new SlimLM small language models, in ...