Your Favourite Chatbots Are Built With Shredded Books

It sounds like something out of a science fiction movie, but it's really happening. Major artificial intelligence companies are buying millions of physical books, ripping them apart, scanning every single page, and then recycling what's left. This process is how they're 'teaching' their powerful AI chatbots, like the ones you might use to write an email or summarise a document.

The idea is to feed these advanced computer programs as much information as possible from real-world texts. The more books an AI can 'read', the smarter and more human-like it becomes in understanding language and generating its own responses. These books become the foundation of the AI's 'knowledge base', allowing it to answer questions, write stories, and even hold conversations.

This method, revealed through recently unsealed court documents, has raised a few eyebrows. While it helps AI get smarter faster, it also brings up questions about copyright and fair use. When a book is scanned and its contents are used to train an AI, is that fair to the original author? It's a bit of a grey area that legal experts and content creators are grappling with right now.

For everyday Australians and small business owners, this behind-the-scenes practice is important because it shapes the quality and reliability of the AI tools we're starting to rely on. Understanding how these tools are built helps us make better decisions about how we use them, and what kind of information we trust them with. It also highlights the rapid and sometimes unexpected ways technology is evolving.