Amazon destroys rare books to train AI models
Company's secret book-scanning operation for AI training revealed
1Amazon's secret book-scanning operation
A 404 Media investigation uncovered Amazon's mass purchase of rare books for AI training. Researchers placed a tracking device in a book they suspected would be acquired by an AI company. The book eventually reached an Amazon warehouse in Nevada. Workers at this facility cut off book bindings for rapid scanning, destroying the original books in the process.
2Details of destructive scanning method
At Amazon's VGT3 warehouse in Las Vegas, employees cut the bindings off incoming book shipments. This method enables fast scanning but results in the destruction of original books. The warehouse's logo features a dinosaur holding a book and baring its teeth, seemingly symbolizing Amazon's approach to books.
3The fate of rare books
The tracked books were rare editions - either very few copies were printed or they were in less common languages. According to a bookseller, these books hold historical, intellectual and sentimental value, though not as celebrated as first editions of Oliver Twist. For AI companies, only the textual content matters.
4Preservation vs industrial scanning
Since 1996, Internet Archive has been digitizing books page by page while preserving their physical integrity. When Google Books launched in 2004, it used industrial-scale but destructive methods. Amazon's current approach appears to be an even more radical version of Google's method.
5Digitizing historical works
Internet Archive developed special machines and software to preserve centuries-old books, creating positions like 'book scanner' or 'scribe operator.' Their method is slower but preserves physical and historical value. Amazon's approach prioritizes speed at the cost of destruction.
6Copyright and ethical concerns
Amazon's practice pushes the boundaries of 'fair use.' The destruction of rare books raises serious ethical questions about cultural heritage preservation. Extracting only text content means losing historical context and physical characteristics of the books.
7The dark side of AI training
Big tech companies need massive data to train AI models, with books being a key source. Amazon's operation reveals the lack of transparency in data collection processes and potential harm to cultural heritage.
8The future of digitization methods
The contrasting approaches of Internet Archive and Amazon spark an important debate about digitization ethics: speed versus preservation? Balancing industrial-scale data collection with cultural heritage protection is becoming increasingly difficult, potentially requiring more regulation in AI development processes.
Original article: Hacker News: Front Page — Amazon nadir kitapları yok ederek yapay zeka eğitiyor