A report by 404 Media found that ISBNdb, a company that tracks book data, offered 'AI training data' from a large book database.

AI companies need data to build better models, and books are a good source because they are well-written and edited.

ISBNdb had a page that advertised a service to hide the identity of AI companies that used their book data, and also offered to destroy the books after scanning.

The company used a process called destructive book scanning, which involves cutting the spine off a book and feeding the pages into a scanner.

Many people were upset by this news, including a financial newsletter that wrote about it.

Elon Musk commented on the post, saying that his company would scan rare books the 'hard way', but didn't define what a rare book is.

ISBNdb has stopped offering the book destruction service, saying it was just a test.

However, other companies may still be destroying books to feed AI models.

There are other ways to scan books that don't involve destroying them.

Training AI models on books makes sense, but it's better to use books that are already online or in the public domain.

Some AI companies have used pirated books to train their models, which is against the law.

A lawsuit found that one company, Anthropic, had used millions of pirated books to train its model.

The company had also destroyed many physical books to scan them, which was revealed in court documents.

A judge found that this was 'transformative' and allowed under fair use laws.

However, the news was bad for Anthropic's public image, with many people criticizing the company for destroying books.

The company had used a machine to cut the pages out of the books, and then scanned them.

This process was very fast, with the company scanning up to 2 million books in 6 months.

This is a problem because it's hard to scan books without destroying them, and many people value physical books.

In the past, book destruction has been seen as a bad thing, and many people are upset by the idea of AI companies destroying books.

There are other organizations, like the Internet Archive, that scan books in a way that doesn't destroy them.

The Internet Archive uses a machine called the Scribe, which turns pages by hand and scans them.

The Archive has scanned many books, but it takes a long time and is not as fast as destructive scanning methods.

The only thing that might stop AI companies from destroying books is public attention and criticism.

This is a topic that involves artificial intelligence and books.