boltTLDR
AI labs are preferring older, pre-2022 books for training data to avoid AI-generated content while building watermarking systems, a shift that torpedoes traditional content-at-scale SEO tactics.
Search Engine Journal reports that AI labs are deliberately purchasing pre-2022 books to train their models while avoiding recently published content infected with AI slop, and are simultaneously developing text watermarking technology. The piece explains why this trend signals the end of content-at-scale strategies as a reliable SEO approach, fundamentally altering how organic search practitioners should think about content production and competition.