← Back to Briefing
AI Companies Resort to Physically Cutting Books for Training Data
Importance: 90/1001 Sources
Why It Matters
This practice highlights the growing, often unconventional, methods AI companies employ for data acquisition, raising significant questions about intellectual property rights, data ethics, and the sustainability of training increasingly data-hungry AI systems.
Key Intelligence
- ■AI companies are physically disassembling books to extract text for training advanced AI models.
- ■This method is used to acquire vast amounts of diverse and high-quality textual data.
- ■The practice likely helps circumvent digital licensing complexities or access content not readily available in digital formats.
- ■It underscores the extreme data demands of current large language models.