Publishers' AI Licensing Negotiations Mark a Change

Driven by the uncertainty surrounding generative AI, news publishers have begun the complex process of drafting licensing agreements that would, in exchange for payment, grant AI firms access to their content. Publishers are faced with the impossible—albeit familiar—dilemma of needing to determine what is best for them now based on what might happen in the future. “It’s essentially an economic question,” said one digital publishing executive, speaking anonymously. “What is the value of the content you have produced, what will its value be in the future, and are you willing to part with that based on what is being offered?” Prior to the advent of gen AI technology, publishers and search engines maintained a largely mutualistic relationship. Now, new search products from companies like Google and Microsoft can answer user questions directly, which could eliminate up to 66% of publishers' search traffic and a huge chunk of revenue, according to estimates from three publishers. In 2023, publishers got 35% of their traffic from search, according to a Press Gazette analysis. But to train and power these products, tech companies need steady streams of accurate, contemporaneous information—the kind that only news publishers can produce. In an October white paper, the News/Media Alliance found that large language models overweigh premium publishers' content by a factor of between 5 and 100. A familiar power struggle The resulting impasse has pushed both parties to explore licensing models in which search platforms pay news publishers for use of their content, akin to the two-year deal struck between the Associated Press and OpenAI in July. The potential arrangements resemble deals publishers have signed with tech platforms in the past, such as when Facebook paid select outlets to license content for its News section. While these deals would differ in their structure and scope, the bargain is a familiar one: In exchange for a lump sum, publishers provide access to their content but risk diluting its value in the process. While the sophistication of gen AI can make it appear unprecedented, in reality, the technology represents the latest iteration of an ongoing power struggle between content creators and content distributors, according to Mitzi Hill, a partner at Taylor English law firm. As such, the shape of the potential partnerships could help define the next era of the internet. Deal structures Publishers working to draft licensing agreements will likely seek compensation for two distinct tiers of data, according to an executive at an ad management firm speaking anonymously. First, they will look for a model that remunerates them for the use of their existing library of content, which gen AI companies have presumably already ingested. The volume of these archives and the pertinence of their information are key variables. Valuing of archival content will be harder to structure, as it can be difficult for publishers to prove that their data has been used, according to the executive., as it can be difficult for publishers to prove that their data has been used, according to the executive. Gen AI search products routinely surface publishers’ language in their responses, but if an article is never regurgitated, it could be hard to prove it was ingested. The second tranche of data, contemporaneous content, should be easier to value, according to Jason Kint, the chief executive of Digital Content Next. This is partially because 10%-15% of daily search queries have never been asked before, according to Google, and AI companies have few other sources of quality, up-to-date information beyond publishers. Publishers could base these license fees on damages, a figure measured by the decline in ad revenue caused by gen AI products answering queries on page, according to the executive. For instance, if a gen AI search product led to a 10% decline in search traffic for a publisher, it could pay a fee to the publisher equivalent to that 10% loss of ad revenue. While such a system could work initially, it would likely evolve, as licensing models typically focus on added value rather than damages, according to Hill. Ongoing legal battles complicate the deals At the same time that publishers are crafting their commercial proposals, others—like News Corp and Axel Springer—are preparing a legal case against AI firms. These arguments typically hinge upon the assertion that AI companies’ use of publishers’ data is not covered under fair use and infringes on their copyright protections, according to Danielle Coffey, CEO of the News/Media Alliance. The simultaneous nature of the two approaches complicates the negotiations, as the outcome of one affects the other. For instance, some legal experts believe that publishers’ copyright arguments will win in a court of law. If so, the likelihood of that outcome could push AI firms to negotiate deals before a verdict is reached, as they would lose leverage afterward, according to Kint. “Tech companies will want to set the market while there is still uncertainty about the legal case and pressure on the IP owners,” Kint said. “Otherwise, they risk waiting for a decision that might weaken their case.” Conversely, publishers opposed to licensing deals have pointed to the strength of their legal argument as the reason they should hold out. “People who are working out how to license their content now, before they have any idea how the landscape will shake out, are working to their disadvantage,” said a second anonymous digital publishing executive. “Negotiating from a position of weakness will get us nowhere.” In reality, there is a strong likelihood that neither publishers nor AI companies will win all of the dozen or so legal cases currently challenging the technology, according to Keith Kupferschmid, CEO of the Copyright Alliance. That makes the situation ripe for compromise. “Publishers and AI companies both think they are going to win,” Kupferschmid said. “No one has leverage, and both have risk. At some point, you have to consider splitting the baby.” This story is part of the Power of Tech Marketing special feature.

AI News

Publishers' AI Licensing Negotiations Mark an Inflection Point

Copyright law favors creators, but commercial compromise offers a hedge against uncertainty

Mark Stenberg

About

Subscriptions

Events

Publications

WORK SMARTER - LEARN, GROW AND BE INSPIRED.

Subscribe today!

Mark Stenberg