After years of scraping and stonewalling, Google is beginning to talk AI licensing with publishers. The shift has triggered a familiar mix of caution and resignation among media execs. They\u2019ve seen this before: overtures of partnership that quietly lead to more platform control.<\/p>\n
The timing isn\u2019t surprising. Amazon\u2019s deal with The New York Times in June<\/a> added urgency to a space where publishers are already uneasy about how their content is being used to train AI systems.\u00a0<\/p>\n Now Google is playing catch up, reportedly holding exploratory talks with around 20 publishers<\/a>. What comes of those talks is still a moving target. But the outlines of what publishers want are already clear. The demands are straightforward, if a little aspirational: real money, real transparency and \u2014 perhaps most elusive \u2014 a sense of control in a market that has rarely offered it.<\/p>\n Digiday spoke with several publishers who broke it down into three core asks: meaningful revenue, clearer terms and a bit more control in a market that rarely gives it.<\/p>\n Revenue that reflects value<\/p>\n This one\u2019s as much about principle as payout.\u00a0<\/p>\n Publishers have spent years feeding the machine \u2014 training the AI, populating the search results, keeping users engaged \u2014 while Google reaped the rewards. Any check now comes too late to feel fair, but it\u2019s still a line in the sand.\u00a0<\/p>\n As one commercial exec at a publisher put it: \u201cYou have to assume the [upfront check] is the last time you\u2019ll see that money.\u201d<\/p>\n That\u2019s why hybrid deals \u2014 flat fees with variable upside tied to usage \u2014 are top of mind. They\u2019ve surfaced in other content licensing agreements, like OpenAI\u2019s deals with the Associated Press and Axel Springer, where payouts reflect not just access but continued use.\u00a0<\/p>\n Without that, the core issue remains: AI Overviews and other products risk turning publishers into invisible suppliers \u2014 content creators cut off from the economics of their own work.\u00a0<\/p>\n \u201cOn day one, it\u2019s fair to say that Google, OpenAI or whoever else it may be won\u2019t be making tons of money [from these days] but eventually they will and publishers should share in that upside over time and not be stuck in some crappy deal where the platform gets all the benefit,\u201d said\u00a0Paul Bannister, chief strategy officer at Raptive, which sells ads for independent sites.<\/p>\n Of course, even a generous revenue share only goes so far. If visibility, attribution and traffic keep slipping, the money won\u2019t matter.\u00a0<\/p>\n Control over how their content is used and surfaced<\/p>\n Beyond compensation, publishers want agency: where and how their content shows up, and under what terms.\u00a0It\u2019s arguably the most important, and least guaranteed, part of these deals. And as they shift from training data to real-time content ingestion<\/a>, that demand is only growing louder.\u00a0<\/p>\n As one publisher, who asked to remain anonymous given the sensitivity of the subject matter, explained: \u201cThe more we can understand how our data is being used by their AI algos, the better we can design going forward.\u201d\u00a0<\/p>\n Until now, that transparency has been limited \u2014 frustratingly so. Publishers say they\u2019re often treated as if they\u2019re asking for something extraordinary when in reality, they\u2019re just asking for basic visibility into how their own content is being used.\u00a0<\/p>\n