ÒÁÈ˾þÃ

Big AI firms have built their models by hoovering up copyrighted material from the internet as training data. They say this is legal, but copyright holders disagree - and this year they hit back in a major way