AI Companies Are Buying Antique Books, Ingesting Their Contents to Train Models, and Then Destroying Them at Incredible Scale

Major AI developers are acquiring rare, old books, digitizing their content for model training, and then discarding the physical copies. This practice highlights a unique approach to data acquisition for advanced AI development.
Key takeaways
- AI firms are using antique books as training data sources.
- Digitized historical texts are being fed into AI models.
- Physical copies of these books are subsequently destroyed.
- This method expands AI's access to diverse linguistic data.
Why it matters
This trend means AI models may soon incorporate historical knowledge and writing styles previously inaccessible. Users of AI tools could see improved understanding of nuanced language and historical context in generated text.
Try this on SynaBot
Related AI assistants, prompts, and tools from the SynaBot catalog.
- ContentsContents — Contents is the Generative AI platform to create impactful content built for performance. It sits in the writing & content category and is built to draft, rewrite, summarize, and structure content across blogs, emails, and documents.
- ContentShake AIContentShake AI, a SEMrush product, assists with generating content ideas, headlines, and outlines. It helps create SEO-friendly content efficiently, accelerating the drafting process.
- Semrush ContentshakeSemrush Contentshake, powered by Semrush, is an AI writer that helps you create SEO-optimized articles and blog posts quickly. It generates outlines, writes content, and suggests improvements based on keyword research. Boost your content marketing efforts.


