Why Are AI Labs Buying Pre-2022 Books?

AI labs are bulk-purchasing pre-2022 printed books to avoid training their models on low-quality AI-generated content (AI slop), while simultaneously developing covert text watermarking technologies. This strategic shift reveals that the mass content-at-scale approach on the internet is unsustainable in the long run and marks the beginning of a new era in search engine optimization and digital marketing dynamics.

Data Quality and Text Watermarking Strategy

To prevent AI models from training on data they generated themselves—a risk known as model collapse—tech giants are turning to human-produced, highly accurate pre-2022 archives. Beyond keeping the AI data pool clean, this move lays the groundwork for watermarking systems designed to trace the origin of future content. For content creators, this indicates that volume-driven mass production tactics can now be easily distinguished by AI algorithms and may turn into a disadvantage.

Industry Reflections and the Future of Content Marketing

The "fast and high-volume content production" strategy long embraced in the digital marketing and SEO world is losing value as search engines and AI models place an increasing emphasis on source quality. While original, thoroughly researched, and human-centric content will stand out in future search engine algorithms, automated generic texts are both being filtered out of AI training sets and facing the risk of penalties from search engines.

*This news report has been prepared based on data published by Search Engine Journal.