A new book looks at the past and future of copyright

From The Economist:

If Walt Disney were still alive, he would be dismayed at the new film role given to his signature character, Mickey Mouse, as a slasher hunting teenagers in an old arcade. The trailer for “Mickey’s Mouse Trap” was released on January 1st. That is the day the copyright of “Steamboat Willie”, the short film that introduced Mickey Mouse’s character in 1928, expired. This early version of Mickey is now in the public domain.

Even before the arrival of a murderous mouse, the field of copyright has been full of dramatic turns, as a new book, “Who Owns This Sentence?”, recounts. That is because “copyright is an edifice of words resting on a long and complicated string of metaphors and double meanings,” write the authors David Bellos, a professor at Princeton, and Alexandre Montagu, a lawyer. Over centuries artists, authors, lobbyists, publishers and public officials have defined and redefined the meaning of copyright, with debate and legal changes happening beyond the public eye.

Initially an author’s right to their creative output was a question of honour. Hermodorus, a student of Plato, published notes he took from Plato’s lectures without permission, which incensed the philosopher. Privileges, a precursor to copyright, emerged in the 15th century in Venice, where authorities gave skilled craftsmen a seven-year monopoly over luxury goods they produced. Among those who profited from Venice’s system was Johannes of Speyer, an apprentice of Johannes Gutenberg, maker of the printing press.

Authors and artists did not always benefit. Powerful publishing syndicates treated books like property and profited from selling the works of Chaucer, Milton, Shakespeare and others. Britain’s Statute of Anne, the first copyright law regulated by government and the courts instead of private parties, came into effect in 1710 and gave authors rights to their works for a limited time (though booksellers and publishers could buy the books before the copyright expired and claim to own the rights perpetually). But it was a decision by the House of Lords in 1774, reasserting term limits for copyrighted works, that fully established copyright in Britain and served as the beginning of modern copyright law in the West.

Mr Bellos and Mr Montagu argue that copyright has gone from a right that favours creators to something more akin to a privilege for the rich and powerful. Two major legal developments led to this. First, in the early 1900s a skirmish over printing posters for a travelling circus redefined the meaning of authorship to include employers of artists. In other words, companies could own copyrights. Second, in 1976 the meaning of literary works was adjusted so that computer software was included.

Intellectual property is now among America’s chief exports. About a third of the 50 richest people in the world derive their fortunes—in whole or in part—from copyrighted goods. Some creators have benefited handsomely (Bruce Springsteen, for example, sold his music catalogue to Sony for a reported $550m in 2021). But many have not. “Copyright is the elephant in the room when it comes to understanding the origins of the wealth gap in modern societies,” the authors write.

Today the newest frontier of copyright is generative ai, which allows people to create images, prose and sounds based on prompts. Artists have called ai “vampirical” because it scrapes pre-existing content for training.

Some writers and artists, such as Sarah Andersen, a cartoonist, have taken ai giants to court over copyright infringement. “We’re not necessarily fighting for ai to go away, but if our work is going to be involved in these systems, we would like to be credited,” Ms Andersen says. She also argues that artists should be compensated and be able to opt out of having their work used to train ai models.

Publishers are in legal tussles, too, including the New York Times, which in December sued Openai (the maker of Chatgpt) and Microsoft for copyright infringement, after talks between the news publisher and tech companies failed. Others are more optimistic. Some news groups, including the Associated Press, have signed deals to share story archives with Openai.

Link to the rest at The Economist

As PG has mentioned on at least two occasions, he believes that feeding images and text into a database that is used to teach an AI is not copyright infringement, at least under US law.

As he understands it, the AI is not capable of creating a copy of a Nora Roberts book even if the words in a Norah Roberts book were used, along with millions of other books, to write new creations that do not include any sentences or paragraphs in any Norah Roberts book or a book/article, etc., by any other author.

The resulting AI cannot create a work identical or substantially identical to a Norah Roberts book. If AI’s creator admits to including the text of a Norah Roberts book along with the text of a zillion other books, how, exactly is Norah Roberts harmed if the AI is incapable of replicating any substantial amount of a book Ms. Roberts has written?

Nobody owns the copyright for, “I love you” or “I hate you,” or anything similar even if an author has included these terms in a million copyright-protected books.