Media Briefing: Publishers confront the AI era during the Digiday Publishing Summit
This article is part of Digiday’s coverage of its Digiday Publishing Summit. More from the series →
This week’s Media Briefing recaps what publishers had to say about AI platforms during the Digiday Publishing Summit’s closed-door town hall sessions.
- ‘We can’t give away our content for free’
- Publishers’ anti-crawler status, Hearst’s OpenAI deal, SearchGPT’s rev-share outlook and more
‘We can’t give away our content for free’
“It’s sort of like the Facebook days.”
That’s how one publisher described the present situation between publishers and generative AI companies in one of the closed-door town hall sessions during the recent Digiday Publishing Summit.
They weren’t wrong. ChatGPT, Perplexity and the like are effectively the new Facebook, Google, etc. But the most apt aspect of the publisher’s statement were the words “sort of.”
The platform era giving way to the AI era offers a do-over — sort of — for publishers. Unlike Facebook, etc., OpenAI and its ilk are making early offers to pay (some) publishers for people to access the publishers’ content through the AI companies’ platforms. Just this week Hearst became the latest publisher to announce such a content licensing deal with OpenAI. Meanwhile, Google has (vaguely) touted the traffic its AI Overviews are sending to publishers, and some publishers are seeing similar referral gains from Perplexity.
But while some publishers may see the skies starting to clear after years of platform referral traffic declines, others are feeling lost in the fog.
“I found that it’s difficult for smaller publishers to get those meetings [to discuss content licensing and revenue-sharing deals with AI companies] when there’s no clear path to contact someone. They’re not typically coming to us,” said a second publisher during the town hall session.
Although the AI companies have struck licensing deals with some publishers, they seem to be unwilling to open their pocketbooks to all.
Case in point: OpenAI’s head of media partnerships Varun Shetty said this week that the platform does not intend to strike ad revenue-sharing deals with publishers for its SearchGPT product and that, instead, publishers will be compensated with “significant incremental traffic from new audiences,” according to Press Gazette.
On a similar note, one publisher during a DPS town hall session said that Google had reached out to publishers that had not opted into its Google-Extended control that permits Google to train its AI models on publishers’ sites. This publisher told a member of Google’s team that it would opt in if the publisher received better attribution, traffic and a share of revenue in exchange. “If it’s only a visibility play, that doesn’t really move the needle,” said the publisher.
And what did Google say to that?
“They just took a lot of notes,” said the publisher. “The person seemed a little bit surprised that we would expect something out of it. We did point to the other conversations we’ve had [in which AI companies would compensate the publisher with credits to use their AI models and/or a share of revenue]. At the end of the day, we can’t give away our content for free. And that didn’t seem to be completely understood.”
Sort of like the Facebook days.
What we’ve heard
“What we have to realize is that we can’t rely on other platforms to send us traffic. We have to create that demand loop itself.”
— Business Insider’s Katie Friedman
Numbers to know
236,000: Site visits that Forbes received via Perplexity’s AI search engine in August.
8: Number of articles, on average, that Reach plc reportedly expects its reporters to file per in-office shift.
159: Number of employees that left Automattic over Automattic CEO Matt Mullenweg’s standoff with WP Engine.
£53.2 million ($69.6 million): How much revenue The Independent expects to generate this year.
0.4%: Percentage share of TikTok accounts followed by U.S. adults that are accounts belonging to journalists, pundits and media outlets.
1%: Percentage share of PubMatic employees that the ad tech firm laid off last week.
What we’ve covered
The platform era for publishers gives way to AI:
- During the Digiday Publishing Summit, publishers discussed the challenges they face from traditional platforms and AI platforms alike.
- How publishers are addressing those two sets of challenges aren’t all that different.
Read more about the platform AI era for publishers here.
What we’re reading
Publishers open up to AI crawlers:
Publishers seem to be reversing course on their efforts to block AI companies’ crawlers from accessing their sites, with the share of popular news sites blocking OpenAI’s GPTBot down from a third to closer to a quarter, according to Wired.
The magazine publisher has joined the ranks of media organizations that have signed content licensing deals with OpenAI, according to Axios.
OpenAI’s rev-share stance for SearchGPT:
Despite signing content licensing deals with publishers, the generative AI company is uninterested in striking revenue-sharing agreements with publishers for its SearchGPT search engine, according to Press Gazette.
Google’s publisher payment ultimatum:
Google is threatening to no longer link to articles from New Zealand publishers in protest to a proposed law that would require tech companies to pay publishers for content distributed via the platforms, according to Reuters.
Known as a newsletter platform, Substack wants to be more akin to Patreon: a platform for creators of all kinds to receive direct compensation from their audiences, according to Semafor.
More in Media
MediaSense buys R3 to strengthen its Asian and North American presence
MediaSense, the U.K.-based media advisory firm, is further expanding its global footprint with the acquisition of fellow advisory firm R3.
Jasper adds new control and marketing knowledge tools for AI-generated content
The Austin startup is adding an API-based tooling to let marketers embed brand guidelines into an AI model for consistent text and visual outputs.
How AI is transforming HR’s ‘dead data’ problem
Recent innovations in AI offer promising solutions for resurrecting that dormant information and converting it into “companion knowledge.”