It’s a slightly strange article because it assumes that an equivalent of “reading” which it posits is permitted under copyright is somehow free. It’s not. If y

It’s a slightly strange article because it assumes that an equivalent of “reading” which it posits is permitted under copyright is somehow free. It’s not. If you want to consume copyright protected content you need - unless the copyright owner allows you - to make some sort of value exchange. Often the primary application of this is that an author (or creator) licenses their content to a publisher (or distributor) and that publisher pays them a royalty for the use of the content. However, the licence that the author provides will (should) specify what uses are covered by the licence. So if an author grants a licence to a publisher then the licence may allow the publisher to print and sell books - but might exclude the right for the underlying literary work to be adapted as a screenplay. So if someone wants to make a film or TV show of that work they need to revert to the author. The publisher may or may not be cut in on that deal. This is a primary tenet of copyright: that the copyright owner can control who does what, and only they can grant that permission, and they can sell (or exchange value) for that permission. It’s a logical leap to say “because the copyright owner has licensed this work to a publisher for people to read it is then ok to use that content to inform composite derivative works using AI”. To expand on the proposal of the article that a search engine “reading” an article is the same as AI “reading” an article: when a search engine “reads” copyright protected content it is creating a signpost - similar to Dewey Decimal signposting - not (generally, but see google news lawsuits) aggregating the works to create derivative works. Addressing the google news point there is an argument that a value exchange has taken place - that the publisher is allowing (or excluding) the use based on the utility they get in terms of increased visibility/audience/ad revenue. If an AI ingests copyright work that is fine so long as the author knowingly permits it and is compensated. However I bet if you said to screenwriters or musicians or journalists “will you licence your work and unique stylistic voice to a company that will create new works that are creatively and stylistically largely indistinguishable from yours” they would say no. Plagiarism is plagiarism and if a human reads and regurgitates a work protected by copyright the copyright owner may (as has been seen in various song lawsuits - and other creative works related lawsuits) have a case to say their copyright has been abused, and be compensated.