Academic Publishers are in a hurry to sign a license agreement with an artificial intelligence company, opening up new revenue streams as U.S. research funds will be cut.
Informa PLC's Taylor & Francis signed a $10 million deal with Microsoft Corp. last year, providing high-tech giant access to training large-scale language models (LLMS) for parts of its library. Bloomsbury Publishing Plc is considering “monetizing academic content through AI transactions.” This is mentioned in the latest result set.
The deal will boost revenues to publish the Giants. Informa sold $75 million of non-repeated data access in 2024 through a partnership between Microsoft and another private partner. Excluding AI trading, organic growth will be close to 3.5%, according to Berenberg analyst William Larwood.
This could offset some of the impact of the US administration's decision to cut federal spending on academic research, Lerwood said in a memo last month. He calculated that around 11% of Taylor & Francis' subscription revenue could be at risk from cuts in US funding in 2026. 43% will be cut to the 2026 National Institutes of Health budget (promoted by US President Donald Trump), leading to a headwind for Taylor and Francis' £33 million revenue. Larwood said a dramatic 25% reduction means a 25% reduction is close to £19 million.
Even those who are eschewing licensing transactions for now, AI is inevitable internally. For example, Springer Nature Ag & Co Kgaa has not signed a license agreement so far, but uses AI to close the workflow. “That's not our number one priority,” CEO Frank Vlanken Peters said in his latest revenue call.
“For us, the main focus of AI is to transform the publishing process and make it faster, more researcher-friendly and more reliable in terms of research integrity,” Springer Nature said in an email response to the question.
The recent surge in licensing agreements in academic circles is part of a broader trend in publishers who are joining forces with fighting strong AI companies.
New York Times Co. recently reached an agreement to use its editorial content over piracy for Amazon.com Inc. after years of court battles between Openai and partner Microsoft. Atlantic and Vox media have signed a similar deal to prevent public companies from scraping the web to train chatbots without providing fair compensation.
“Licensing transactions could be the easiest way to resolve copyright disputes between developers of large language models and owners of content that these models are trained.”
Compensation remains a key point of competition in AI trading for academic publishers. Some authors whose work has been approved have not been given the opportunity to opt out of partnerships and have not been properly compensated for.
Microsoft offered HarperCollins $5,000 per title, splitting 50/50 with the author. According to Mary Raysenberger, CEO of the Author's Guild, there is no ballpark when it comes to what writers can expect. That position should be asked to authorize and be given 75% to 85% of the revenues from AI transactions. The author was recently paid just $97 by Taylor & Francis to give him a book on LLM training, she told Bloomberg.
“Revenues from these partnerships are shared with the authors and other rights holders in accordance with the terms of the license and the term of the loyalty statement in the agreement,” a spokesman for Taylor & Francis said.
“Scholars don't expect to make that much money from books,” says Rasenberger. “They don't want to lose control of their work.”
Beyond the sour relationship with authors, greater concern for publishers is becoming obsolete in the world of AI-generated content.
“It's definitely a risk,” said John Davis, an analyst at Bloomberg Intelligence. The key question, he added, “How can we give the authors and shareholders certainty that they are not driving out the baby with bath water?”
The publisher's experience with data sales rights should be important in implementing the right guardrail, Davis said. Taylor & Francis negotiated clear rules for display rights, reuse, and counting adjacent words (a verbatim extract of more than a certain word).
Belenberg's Larwood isn't too worried, either, referring to intellectual property rights built into the Taylor & Francis agreement. “This suggests that we will appear to sign more contracts in the future, and we'll turn upside down on the estimate,” he said.
This article was generated from an automated news agency feed without any text changes.
