EmbeddingGemma 2: an open, lightweight multimodal embedding model
deepmind.google·6d ago
AI Engineering
Hacker News·Jul 31·5 min read
AI companies destroy rare and non recoverable physical books
TL;DR
AI companies are increasingly engaging in 'destructive scanning' of rare physical books, cutting them apart to create training data for models. This practice raises ethical concerns as millions of unique volumes are destroyed in the process.
✦ Why It Matters
Consider advocating for the preservation of physical books and ethical data sourcing in AI.
Key Takeaways
- Be aware of the impact of AI on cultural heritage.
- Advocate for legal changes to protect rare books.
- Support local booksellers facing bulk purchase requests.
- Engage in discussions about ethical AI practices.
How It Works
Companies like ISBNdb buy rare books, cut off their spines, scan the pages, and then shred the originals. This process is marketed as a form of digital preservation, despite the irreversible loss of the physical book.