The plot thickens… HarperCollins became the first Big Five publisher to sign an AI licensing deal, confirming it reached an agreement with an unnamed tech company (spoiler: it’s said to be Microsoft). Harper would let Microsoft use some of its nonfiction titles to train AI models — but only if authors opt in. Authors who agree to have their work hoovered up for training purposes will reportedly get a $5K fee per title, which they would split evenly with Harper. Microsoft, which has integrated AI across its products, reportedly wants the books for an upcoming model.
May AI use your content?… As chatbot copyright-infringement lawsuits pile up, AI companies are starting to strike major content-licensing deals. HarperCollins parent NewsCorp signed a reported $250M agreement with OpenAI this year that would let the ChatGPT maker use articles from its pubs (including The Wall Street Journal and Barron’s) to answer queries and in training. OpenAI has made similar deals with other news publishers including Axel Springer (Business Insider, Politico), Vox Media, Dotdash Meredith, and Time. And Microsoft partnered with Reuters, Hearst, and Axel Springer on AI news features. In September, “John Wick” studio Lionsgate gave AI-video startup Runway access to its movie library.
Playing defense: The New York Times and others have sued AI companies, alleging they stole their work for training. There are also several suits from authors including George R.R. Martin, John Grisham, and Jodi Picoult.
If you think you can’t beat ’em, join ’em… AI companies are trying to avoid legal fallout before court battles set a precedent, but publishers are also trying to cover themselves in case the gavel doesn’t fall in their favor. Authors Guild CEO Mary Rasenberger said: “AI is here to stay. Our goal is to make sure AI use becomes part of a licensing regime rather than to be considered fair use.”