Putting the insane logistics of actually identifying all of the copyright owners of various images/articles/content/etc. and then paying them aside...

Do you think it's fair for the AI companies to effectively use all of the world's published data for free?