At Pan Macmillan, Sara Lloyd has been appointed to the new position of group communications director and global AI lead. Communications and publicity directors of Pan Macmillan’s imprints will report to her, and she will coordinate on AI strategy with all of the company’s English-speaking divisions worldwide. Llyod “will chair a newly established global trade AI steering group, coordinating the group’s approach to AI safety and ethics, and will develop cross-group principles and policies in line with AI developments,” the Bookseller reports. Lloyd said, “As part of the Holtzbrinck Publishing Group, the impetus to lean into new technologies, innovate and […]
AI
Author Sues OpenAI and Microsoft on Behalf of Nonfiction Writers
Julian Sancton, author of Madhouse at the End of the Earth: The Belgica’s Journey into the Dark Antarctic Night, has filed suit against OpenAI and Microsoft on behalf of other nonfiction authors, citing copyright infringement in training ChatGPT, Bing Chat, and other generative AI tools. Unlike other similar suits, this one focuses on the work of nonfiction authors, and is the first to name Microsoft as a defendant. Like other suits, this one asserts that the tech companies are making a profit—the complaint quotes an OpenAI report that it is making $1.3 billion per year—from the use of copyrighted works, without […]
Saying the Quiet Part Out Loud On Generative AI
As we have noted recently, the various AI companies (and their backers) have dutifully filed comments with the US Copyright Office explaining why they believe training their systems on copyrighted work is fair use and “just math.” So it’s notable that vice president of audio at Stability AI Ed Newton-Rex left the company recently and made clear at the Cerebral Valley AI Summit he believes those arguments are wrong. (He oversaw their music creation product, Stability Audio.) He resigned “because I don’t agree with the company’s opinion that training generative AI models on copyrighted works is ‘fair use’.” As he […]
Judge Dismisses Portions of Authors’ Lawsuit Against Meta
Judge Vince Chhabria of the US District Court in Northern California has dismissed portions of Richard Kadrey, Sarah Silverman, and Christopher Golden’s class action lawsuit against Meta. The judge granted Meta’s motion to dismiss claims that the output of Meta’s large language model, LLaMA, infringes on their copyrights, saying that he didn’t understand how the AI tool’s output was similar to Silverman’s book. The plaintiffs may amend the claim, but for it to stand they must “argue that LLaMa’s output was substantially similar to their works,” Reuters reports. Meta has not responded to the authors’ main argument, that their books’ […]
While Copyright Office Considers, FTC Is Already Concerned that Generative AI Is Unfair Competition for Creators
As the US Copyright Office considers the thousands of comments on “copyright law and policy issues raised by artificial intelligence systems,” the Federal Trade Commission drew attention with a press release to their own position on generative AI in the marketplace — suggesting the agency stands ready to protect creators from unfair competition and consumers from non-transparent machine-generated content. From a pure copyright perspective, resolving issues around fair use could take months to formulate and legislate, or more likely years to litigate, while the world changes before creators’ eyes. Importantly, the FTC states more definitively what creators have been waiting […]
Copyright Office Shares Comments On AI Legal Issues
Earlier this year the US Copyright Office solicited comment on “copyright law and policy issues raised by artificial intelligence systems,” in order to “help assess whether legislative or regulatory steps in this area are warranted.” They have received nearly 10,000 comments, which were recently posted into a searchable archive, including extensive submissions from market leaders such as OpenAI, Google and Meta. Google takes the classic tech company approach, suggesting any issues should be left to the courts (where final adjudication will be glacial). Like all of their peers, they argue that large language models (LLMs) are just math; they “capture […]