Post by : Jyoti Gupta

Photo : AP

A big tech company used millions of books to train its AI without asking the authors first. A judge in the US said this was okay because the company didn’t copy the books to steal from the writers but to help the AI learn and create new things.

The judge compared it to someone learning to write by reading other books—not copying them, but using them to get better. That part is called "fair use", and it’s allowed by US law in some cases.

But there was a problem. The company also downloaded and saved over 7 million pirated books from the internet and stored them in one place. The judge said this wasn’t fair use and broke copyright law.

Now, there will be another trial in December to decide how much money the company has to pay for keeping those pirated books. Even though they later bought many of the books legally, that doesn’t cancel the earlier mistake.

The case is important because other AI companies are also being taken to court for using books, news articles, and other content without permission. Some people think this will one day go to the Supreme Court to make a final decision for all.

This court decision only applies in the US. In other places like the UK, the rules are different, and it’s harder to use copyrighted content without permission.

Generative AI