Media Briefing: Publishers debate the value of AI licensing and GEO

To be expected, AI was a popular topic during the Digiday Publishing Summit in Vail, Colo. last week. In the previous edition of the Media Briefing, we recapped publishing execs’ thoughts on the impact of referral traffic declines from traditional search platforms on their businesses, due to the growth of AI-generated search summaries, and a renewed push to build direct reader revenue from subscriptions and memberships.

But during the summit’s town hall, execs also talked about the lack of traffic coming from AI search platforms, and debated the value of chasing after AI licensing deals and optimizing content to improve their visibility in AI search (often referred to as GEO) if they were receiving very little traffic in return – and revenue from AI licensing marketplaces was still hard to come by.

“For publishers, our appearance [in AI answer engines] is not going to drive traffic,” said one exec. “It’s all about partnerships.”

Publishing execs spoke under Chatham House rules so Digiday could share what was said while maintaining the executives’ anonymity

Here’s what publishing execs had to say at the Digiday Publishing Summit about AI licensing deals and marketplaces, and GEO:

Weighing the value of AI licensing deals and marketplaces

One publishing exec argued that AI content licensing was pivotal to protecting content, and gives a publisher leverage for potential future lawsuits.

“If you’re not doing licensing, you’re basically telling the world that it’s okay to take your content and if you ever decide to sue, you will have nothing really to base it on because you stop protecting it. If you’re smaller, you can participate in a collective or some type of consortium. That’s helpful. There are also a lot of brokers out there that will support smaller publishers if they don’t have a licensing arm,” they said.

Another exec echoed this sentiment that AI licensing deals were a way for the publisher to get money from tech companies, but also a defensive move to protect their content from getting scraped by LLMs without compensation. 

“We have a whole team dedicated to partnerships and licensing, and we do make money off of them. But it’s more like, we want to protect our users and our data, and [our content],” they said.

Other execs, however, weren’t so hopeful. 

“I’m slightly dismissive of how much revenue you’re going to get from these licensing deals. But please, someone tell me if I’m wrong,” one publishing exec said.

Skepticism aside, publishers know they can’t afford to ignore this. One publishing exec said their business development team was focused on signing smaller AI distribution deals (such as with ProRata), in order to test these partnerships before seeking larger deals for more revenue.

“We’re not necessarily looking at it from a revenue standpoint at this point. We’re looking at deals that are a little bit more publisher-friendly as they’re coming out with more publisher programs across the industry,” they said. “We also look at operational lift. [If] it’s less of a lift, then it gives us this opportunity to test, even though they may not be at scale at this point.”

But the exec added their ambitions are growing. The publisher is now scrutinizing any AI licensing deal that comes across their table, such as the AI licensing marketplaces being developed by companies like Microsoft. These deals are evaluated against guardrails. Those included not allowing deals that included archival content, how much editorial control the publisher would retain, how long the AI platform could keep the content, the ability to remove content if it was being misrepresented in the AI platforms, attribution of content, and ultimately, revenue.

Execs also discussed the growing trend of publishers selling insights into how AI search engines surface their content — and when they’re being cited in AI-generated summaries — to develop and sell GEO products to clients as part of branded content packages.

One exec said branded content wasn’t the only route. 

“if you’re going to be partnering with some retailer or some brand and they’re not appearing [in AI answer engines] as much as they want to, or they want to appear more in answer engines, then you can work with them to identify the content that’s already on your site that is not GEO-optimized for answer engines, and then optimize it for it, and then you can also write supplementary content around that – new content,” they said.

Who’s handling AI licensing deals?

The town hall also surfaced a practical question: who inside publishers is actually responsible for pursuing and evaluating AI licensing deals, and how much internal resource is being devoted to it. 

The range was striking. One publisher had around a dozen people working on it in a fractional capacity. Another had just two part-time people. One exec said their company had fewer than 10 people focused on AI licensing specifically, while another had roughly 10 people covering all licensing, including more traditional deals with tech platforms and news aggregators like SmartNews and SmartBreak. One head of business development said they were leading a team of four, but that they were primarily responsible for AI licensing deals and were spending about a quarter of their time on this.

“The AI stuff is more complicated because you don’t know what you don’t know. It’s all in the language, in the contracts. The tech companies are just so good at wanting to get [everything] from us, and it’s a complete struggle to fight with them to do that. So the AI stuff is much harder,” another exec said.

What we’ve heard

“We’ve done so much work on our pages to clean up the experience. We’ve taken away a lot of ads on our page because we wanted people to focus on the call to action and actually convert. So now on our commerce pages, we have less than two ad spots… We’ve done so much work to improve clickthrough, and what we’ve seen in the last year is that clickthrough has actually decreased… It’s not the narrative that we were expecting to see.”

— A publishing exec during an editorial roundtable session at the Digiday Publishing Summit in Vail, Colo. last week.

Numbers to know

27 percent: The percentage decrease in Business Insider’s subscriber base from 2022 to 2025.

20 percent: The year-over-year decline in Future plc’s online audience in the first half of its fiscal year. The publisher blamed Google referral traffic challenges.

600: The number of articles a Fortune editor has written using AI since last July.

Nearly 20 percent: The percentage of Fortune’s web traffic in the second half of 2025 driven by AI-assisted stories.

80: The average words typed per minute by New York Times journalists, double the average typing speed in the U.S.

What we’ve covered

Publishers see double-digit growth from The Trade Desk’s OpenPath, but volatility remains

• Publishers say OpenPath revenue has largely held up – give or take a few sharp dips and recoveries in recent months.
• Nine publishers told Digiday their OpenPath revenue and yields have been strong over the last six months, spurring some of them to trial OpenAds.

Read more here.

How Dow Jones, Forbes, The Guardian and other publishers’ revenue streams are shifting in 2026

• Many publishers have accelerated efforts to diversify their revenue streams through ads, events and bundled subscriptions. They’ve also started to shift search strategies.
• But ads are still dominant, subscription growth is not a given, and events and newer streams like podcasts and newsletters are gaining ground, according to the latest Digiday+ Research report.

Read more here.

Time pitches GEO insights into a new brand offering

• Time is taking what it’s learned about where AI search engines pull information and when Time shows up in those summaries to sell a new GEO product for brands, Mark Howard, Time’s chief operating officer, announced onstage last week at the Digiday Publishing Summit.
• Time’s new product focuses on brand sentiment, analyzing how AI search engines talk about a brand, pinpointing the potential mismatch between what a brand wants those AI summaries to say about it, and then selling Time-branded content to those brands to improve their messaging.

Read the recap of this DPS session here.

Why The Guardian’s first reader-facing AI product isn’t a chatbot

• The Guardian didn’t want to build an AI chatbot, concerned that it would risk that chatbot misrepresenting the news publisher’s journalism and undermining readers’ trust – so it built something else, Chris Moran, head of editorial innovation at The Guardian, explained onstage at DPS.
• The Guardian has begun to roll out its first reader-facing AI product, called Storylines, an AI-generated spin on the “related links” module common to publishers’ pages.

Listen to the latest Digiday Podcast episode, recorded live at DPS, here.

After newsroom cuts, The Washington Post turns to creator-led video deals 

• The Washington Post’s creator network has debuted its first creator-led video series, opening up a new revenue stream as it looks to keep content costs down following recent devastating newsroom layoffs.
• Central to the approach is a key distinction: creators hired by the publisher retain ownership of their IP, according to Sara Kehaulani Goo, president of The Washington Post’s creator network. That’s a more economical model for the publisher than retaining the IP itself. 

Read the recap of this DPS session here.

Inside The Daily Mail’s creator-led content playbook

• Daily Mail publisher DMG Media has hired more than two dozen creators since it launched its creator-led social channels in October 2025. 
• So far, creator-led content has generated on average between 250,000 and 300,000 views, which is above the average of DMG Media’s traditional social channels (Daily Mail and Mail Sport). They’ve also attracted branded partnerships and sponsorship opportunities.

Read more here.

What we’re reading

AI is finding its way into opinion journalism, often without disclosure

Major newsrooms are publishing opinion pieces that seem to include text generated by AI, often without clear disclosure to readers, The Atlantic reported. 

Tech reporters are also increasingly using AI to draft and edit stories

Independent tech reporters are also increasingly using AI to help write and edit stories, to make up for fewer editors and fact-checkers, according to Wired.

Versant is looking at buying Vox Media’s podcast business

CNBC and MS NOW owner Versant is one of multiple companies in talks to buy Vox Media’s podcast network, The New York Times reported.

Legacy media is chasing the creator playbook

CNN and The New York Times are leaning into podcast-style video and creator playbooks to win back audiences, according to Status. (The New York Times recently hired a video training editor to help reporters and editors learn more video skills.)

New York Post “runners” keep one of the oldest jobs in journalism alive

Here’s a fun one, from Semafor: A day in the life of one of New York Post’s “runners” (or on-the-ground, general assignment reporters that chase the news of the day), who embody the survival of shoe-leather journalism at a time when AI tools are increasingly automating and aggregating newsroom workflows.