---
title: "How publishers plan to tame AI’s Wild West"
id: "12993"
type: "post"
slug: "how-publishers-plan-to-tame-ais-wild-west"
published_at: "2026-11-12T13:00:00+00:00"
modified_at: "2026-11-11T22:54:08+00:00"
url: "https://mediacopilot.ai/how-publishers-plan-to-tame-ais-wild-west/"
markdown_url: "https://mediacopilot.ai/how-publishers-plan-to-tame-ais-wild-west.md"
excerpt: "David Buttle says the market for AI's use of news is still lawless. SPUR is betting that collective action can change that."
taxonomy_category:
  - "AI media analysis"
taxonomy_post_tag:
  - "publishers"
  - "google"
  - "AI licensing"
  - "media"
  - "AI media"
  - "AI search"
  - "SPUR"
  - "Grounding"
  - "David Buttle"
---

# How publishers plan to tame AI’s Wild West

David Buttle says the market for AI’s use of news is still lawless. SPUR is betting that collective action can change that.

It’s been a strange couple of weeks for anyone tracking how AI companies use the news. On Nov. 3, an adviser to the EU’s top court said chatbots that pull articles in real time to answer questions may, in some circumstances, infringe publishers’ rights. A few days later, Meta paused a set of new business terms that publishers had attacked as a license to crawl their sites for AI. Neither is a final verdict on anything. But both point the same way, and both happened because publishers pushed back together.

That last part is new. For most of the past 20 years, the media industry has been terrible at collective action. Every publisher cut its own deal with Google and Facebook, and the platforms were happy to keep it that way.

**Listen or watch:**

- [Watch on YouTube](https://www.youtube.com/watch?v=YmC7iLybRoY)
- [Listen on Spotify](https://open.spotify.com/show/3UAlPpLmb9GXpdmXMIVTDd)
- [Listen on Apple Podcasts](https://podcasts.apple.com/us/podcast/the-media-copilot/id1712909951)

David Buttle thinks AI is the moment that changes. David runs DJB Strategies and spent years at the Financial Times, where he was involved in the paper’s deal with OpenAI. He’s also co-lead of SPUR, the Standards for Publisher Usage Rights coalition, which launched in February with the FT, the Guardian, the BBC, Sky News and the Telegraph and now counts well over 100 members. On the latest episode of The Media Copilot podcast, I asked him what it’s actually for.

### The supply chain argument

David’s pitch is about control. In print, publishers owned their supply chain end to end. In the digital eras of search and social, they kept making the journalism, but the platforms decided how it reached people and how it made money. AI is rewiring that supply chain again, and David’s view is that a market which is still forming can still be shaped.

“There is an opportunity for publishers at this moment in time to come together and exert more control over their supply chain,” he told me.

The flip side is the threat. Publisher content is being taken at industrial scale and turned into products that substitute for the original. David breaks the leakage into three buckets: Google tying AI use to search, declared bots that ignore publisher signals because a human triggered them, and outright infringement by scraping services. SPUR’s strategy is to make the cheap ways of getting content more expensive, so that paying becomes the rational option.

The catch is substitutability. If one publisher blocks a scraper, the scraper just goes to a competitor. That’s why this only works when whole segments of the industry move together, which is the whole reason for a coalition.

### Measure first, then price

The most concrete thing SPUR has produced is its content telemetry standard, which defines five things that can happen to a piece of content inside an AI system: it gets retrieved, grounded, cited, displayed and engaged with.

Grounding is the one that matters most. Retrieval is grabbing a pile of 50 articles. Grounding is the handful of passages the system actually uses to build its answer. A publisher whose work grounds an answer but never gets a citation has just lost a reader without anyone telling them.

That’s the problem with Google’s AI payments pilot, as David sees it. The data publishers get covers citations and clicks, but not grounding, so they can’t tell how much value they’re giving up. He called the pilot “a small scale test that is effectively a hedge” against Google’s preferred outcome, which is not paying for content in AI features at all. Harsh, maybe. But if you can’t measure what’s being used, you can’t put a price on it.

- Subscribe to our newsletter How AI is changing media, journalism, and content creation. [Learn More](https://mediacopilot.ai/newsletter/)

![Image](https://mediacopilot.ai/wp-content/uploads/2025/05/ads-1024x682.jpg)

### Ads for bots, and other side bets

I’ve been interested in the idea of serving ads directly to AI crawlers, which TIME is doing with Mobian. David was diplomatic about it: good luck and bravo, as he put it, but he doesn’t see it replacing human ad revenue. AI assistants are built to give users the best answer, and an answer somebody paid to place in front of the bot works against that. He also doubts the AI companies would want these ads to work given what it would do to user trust.

I’m not fully convinced it’s that simple, since ads can borrow a publisher’s reach in a way that’s at least interesting. But I take his larger point. Bot ads look more like an arbitrage play than a business model, and they could give AI companies a convenient excuse to skip paying for the content itself.

### What a small publisher should do

David’s advice for a small newsroom with a deep archive was the most practical part of the conversation: focus on what you control, and don’t wait for a court case to save you. Look at your traffic and the user needs you serve, invest in the kinds of content AI can’t easily replace (analysis, opinion, niche expertise), and treat AI access as a dial rather than a switch. Older or commodity content can stay visible to AI systems to keep your presence there, while the work that drives subscriptions stays protected.

None of this adds up to a functioning licensing market yet. Outside the biggest publishers, what comes back from AI today is mostly visibility, with very little cash. David is clear-eyed about that. What worries him is how many publishers won’t make it through what he calls the Wild West era of AI.

The coalition’s bet is that collective pressure can shorten that era. The Meta reversal suggests it can work. Whether it works fast enough is the real question.

**This episode is sponsored by Genna and Yoast.** [Genna](https://genna.com/?utm_source=mediacopilot&utm_medium=podcast&utm_campaign=mediacopilot_nov26)
 turns written content, interviews and recordings into publish-ready video in minutes using real, licensed footage. Yoast SEO AI+ shows how ChatGPT, Gemini and Perplexity talk about your brand; [join the Yoast Black Friday waitlist](https://yoast.com/black-friday?utm_source=https%3A%2F%2Fmediacopilot.ai%2F&utm_medium=podcast&utm_campaign=media-copilot-promo)
.

### In this episode

Pete and David discuss:

- From “AI-resilient content” to the zero-click future
- Why publishers needed SPUR now
- Whether a real market for AI content is forming
- Three ways content gets taken without fair payment
- Blocking bots, and why it has to be collective
- SPUR’s content telemetry standard, and what “grounding” means
- Google’s AI payments pilot, and the data it leaves out
- Advertising directly to bots, and whether AI companies want it to work
- Courts, fair use and real-time AI
- Advice for a small newsroom with a big archive
- What keeps David up at night

### About the guest

**David Buttle** is founder and CEO of [DJB Strategies](https://www.djbstrategies.com)
 and co-lead of SPUR, the Standards for Publisher Usage Rights coalition. He was previously the Financial Times’ director of public affairs and platform strategy.

**About the show:** To explore more conversations like this and see what’s new, visit the Media Copilot website at [mediacopilot.ai](https://mediacopilot.ai)
. You’ll find new episodes, expanded resources, and tools designed for journalists, communicators, and media leaders working through the fast-changing world of AI. It’s the home base for everything Media Copilot and it’s just getting started.

**Enjoyed this episode?** Subscribe to The Media Copilot on [Substack](https://mediacopilot.substack.com/)
, [Apple Podcasts](https://podcasts.apple.com/us/podcast/the-media-copilot/id1712909951)
, [Spotify](https://open.spotify.com/show/3UAlPpLmb9GXpdmXMIVTDd)
, or your favorite app. On YouTube? Tap the Like button and Subscribe to the [YouTube channel](https://www.youtube.com/@themediacopilot)
.

Produced by The Media Copilot.

Music: “Favorite” by Alexander Nakarada, licensed under CC BY 4.0

All rights reserved. © AnyWho Media 2026

## TRANSCRIPT

**[0:02] Pete Pachal**

Hi, welcome to the Media Copilot. It’s a podcast that’s all about how AI is changing media, news, and communication. My name is Pete Pachal. I covered tech for a long time as a journalist, and now I have conversations with the media people, the builders, and the creators who are all working to answer the question how will we get information in the future? And what does that mean for the people whose business is information, especially media?

**[0:28] Pete Pachal**

For about 20 years, publishers have been pretty bad at acting together. Every company put its content on Google and Facebook and then got dependent on them, and the platforms liked it that way. AI might be the thing that finally changes that. AI systems are taking publisher content at industrial scale, and the market for paying for that content is really still more like the Wild West than a marketplace. My guest today is trying to change that. David Buttle is the founder of DJB Strategies and one of the people behind SPUR, the Standards for Publisher Usage Rights Coalition, which has grown from five British publishers in February to more than 100 members around the world. We talk about why publishers needed SPUR now, what its new standard for measuring AI’s use of content actually tracks, why Google’s AI payments pilot looks like a hedge, and whether advertising to bots can make up for lost human traffic. Of course, we also talk about what publishers should do right now, especially smaller ones.

**[1:35] Pete Pachal**

So before we get to that, quick note if you’re listening on Apple or Spotify, it would be great if you could leave a rating or a review. And if you’re watching on YouTube, please like the video and subscribe to the channel. There, I think there are a couple of buttons down below there. Those things really do help more people find the show.

**[1:54] Pete Pachal**

Okay, let’s get going and to my conversation with David Buttle.

**[1:59] Sponsor messages**

Genna and Yoast.

**[3:29] Pete Pachal**

David Buttle, welcome to the Media Copilot.

**[3:32] David**

Thank you, Pete, it’s great to be here.

**[3:35] Pete Pachal**

Yeah, it’s kinda welcome back ’cause we did that webinar a bit ago. that that did go out in the feed, but this is it. This is the first proper media copilot podcast episode you’re doing.

**[3:46] David**

Amazing, yeah, yeah, yeah, that was brilliant. And yeah, I’ve been looking forward to this.

**[3:51] Pete Pachal**

Yeah. me too. Like you’ve been busy since we last, I guess, formally chatted. so ever s back back then we were talking about publishers and AI’s resilient content, sort of what kinds of content AI would substitute easily versus stuff that was like maybe a little more protected. and now like the c I feel like the conversation has evolved considerably. everyone is acutely aware of what AI is doing to our media ecosystem, which is sort of, you know, rapidly approaching what people call the zero click future, right? The traffic’s going off a cliff. AI is clearly you know, at the core of this. and you have been very active in what what seemed like a unattainable thing a couple years ago, which is like collective action on like what are we supposed to do about this? And you’ve you’ve form this this group called SPUR, right? And I I would love to hear more about that, like how those conversations you were having a year, year and a half ago, you know, turned into this this push and then this organization, which is, is rapidly growing. So like what what convinced you to can not just you to con that publishers needed SPUR, but what convinced them to join something like this?

**[5:15] David**

So I think, I mean, this is a kind of unique moment in time, or at least one that we haven’t seen for 20, maybe 30 years. It was apparent to me, it’s apparent now, that there is a new distribution paradigm that is in the process of being, of playing out through the market. And there is an opportunity for publishers at this moment in time. to come together and exert more control over their supply chain than they’ve had in the previous distribution eras. If we think about print, they had very good alignment and ownership of their supply chain. They were able to create shared incentives to maximize their prominence, their volume sales. And in the digital eras, those that fragmented, they still employed the journalists, they still the journalism, but the product manifestation of that is heavily influenced by the platforms and they rely on the platforms to distribute and to monetize. And AI, it’s apparent is going to change that supply chain for information in a lot of circumstances. So the kind of central idea of SPUR is that if publishers come together at this moment in time, as this new market is forming, then their likely to be able to place themselves in a better strategic position and with more control over their supply chain than they’ve had in the areas of search and social media. So that’s the kind of central strategic. principle, an idea that sits behind SPUR. And I think when we articulate that to publishers, then they see it. That’s sort of painting the opportunity. But there’s obviously also a downside risk, which I think sharpens the focus. And that’s the fact that publisher content is now being taken at industrial scale by a new industry that has emerged. And that is then used to create directly substitutional

**[7:26] David**

products and that is eroding value from publishers even further. So there is this upside opportunity to shape the rules and the norms and the standards around which this new market forms but also there’s this increasingly acute downside risk of substitutional products being created off infringed works and unauthorized access.

**[7:49] Pete Pachal**

Yeah, because that’s my natural question, as you were mentioned a couple of times that this new market’s forming, which is is a new market forming? You know, and I know it’s been a few, few years already, but it also always seems like we’re we’re kind of waiting for something to really catch fire. Obviously, there’s a a bunch of solutions that are still in the market, some sort of further along than others. You know, we could go over them, like, you know, there’s licensing, there’s sort of paper crawl, paper use type business models, there’s even new emergent ones that involve advertising directly to bots, you know, as bot traffic increases, as humans decreases. So where where where do you see the energy around all of the sort of ways you can monetize? And is, you know, how does how is Spurs sort of forming around these sort of various I guess alternatives if if you can call them that?

**[8:47] David**

So yeah, there is a market forming. There’s a market forming sporadically on the permissioned route, largely via bilateral deals or deals which are primarily around paying off legal risk. And then…

**[9:05] Pete Pachal**

You’re talking about the big sort of licensing deals between sort of the big AI companies and the more mostly the major publishers, right?

**[9:12] David**

Exactly, Yes, precisely that. And then there’s a more liquid, vibrant market emerging, which is not centered on publishers, but centered on scraping services that gain access to content on behalf of demand side platforms and then sell it on. What SPUR exists to do is to create the economic incentives such that the value that’s going into that market instead goes to the organisations that pay for the journalists that create the content in the first place.

**[9:47] Pete Pachal**

And how do you do that? How do you in create those incentives? Cause we’ve definitely all read about at least, you know, media nerds have all read about the that sort of s central piece of all these, you know, what the the companies that are scraping on an industrial scale and then sort of reselling the content to their own customers. And you know, I kind of call it a gray market sometimes because some seem a little more interested in the publisher side of things than others. So presumably they would respond to certain incentives. But you tell me the state of that and it seems like, you know, we t think about that as kind of like a big industry, but it seems like there’s a little more texture there, right?

**[10:27] David**

Yes, so the way you, I mean, we look at this through an economic lens. So at the moment, there are basically three ways that publisher content is taken and used without fair payment. looking slightly reductively. The first is by Google, who have historically tied the use of content for AI to the use of content for search. The Competition and Markets Authority in the UK has created new rules for Google, and Google is responding to that with changes in its product design. But that’s the first sort of vector of value loss. The second

**[11:04] Pete Pachal**

Yeah, you can now opt out of AI, right? Like that was the thing that you couldn’t really do before. You had to sort of be in overviews to be in search, as I understood it.

**[11:12] David**

Yeah, exactly right. So you can opt out of AI, but those AI summaries are sort of they’ve been bundled together into a single opt out and then deployed very aggressively in certain parts of Google’s estate, specifically Google discover. So the more traffic is mediated via a discovery, a discovery summary, an AI summary, which makes that opt out decision harder. So Google can protect summaries where it really matters to Google, which is around search because it’s in this competitive fight with chat GBT. The second way that publisher content is taken without fair payment is by declared bots, but bots that choose to ignore publisher signals or a program to ignore publisher signals on account of being triggered by a human. Now this is going to be more concerning as agentic systems sort of diffuse through the consumer base.

And then the third way are infringements, outright infringements, from sort of clandestine scraping services. And what SPUR is trying to do, and what SPUR is working to do, is to raise the economic cost associated with the last one of those and to talk directly to the platforms around the first two of those in order to force the licensing market into existence, mean force the value that’s flowing into this market through the pipe that goes to publishers instead of into scraping services. So that’s at a high level how Sperth thinks about this. Sperth is a technical organization and as I say, thinks about this in economic terms. Technically, we’re never going to be able to be completely watertight to threats, IP threats, but what we can do is combat the low cost ways that publisher content can be accessed without permission. And if we combat those effectively enough, then we’ll push the price up above the price of the licensing market. And at that point, the economics tell us that we’ll have a licensing market in which publishers writ large can participate. So that’s how we’re thinking about this problem. Alongside that,

**[13:27] Pete Pachal**

Is that how is that going, if you don’t mind like in terms of like do you must be looking at metrics that sort of are pushing that economic cost to sort of just taking content? I actually just the the last person I spoke to, I spoke to Jonathan Roberts recently, at People Inc, who and those guys do better than anyone, as far as I can tell, in terms of blocking bots. And they do they they’re fairly successful at it, although they’re not like you said, it’s never a hundred percent watertight. but To your point, it doesn’t really need to be. Is that something? But I I always sort of say I’m not sure if that’s a good sign or a bad sign, because not every publisher is the resources of a people ink, right? but they do have sort of some support from groups like yours, from CDNs like Cloudflare. So it feels like directionally things are going, but you would know better than I would on like sort of the gap with that economic gap, like when it becomes you know, cost prohibitive for people to scrape without compensation.

**[14:23] David**

So there’s a challenge here around collective action because there is a degree of substitutability across publisher content. If you can’t get People Inc’s content, then perhaps you can get it from one of their competitors. So we need segments of the industry to put these protections in place together. So that’s how we look at this problem. We’ve run very extensive tests across cohorts of ASPR members to understand

**[14:51] David**

What are the ways, the low cost ways that the content is being accessed now and how can publishers improve that? think directionally, absolutely, I think we will look back at this era that we’re in now as the Wild West, like the early days of digital advertising. So I think we’ll get there, but what SPUR is seeking to do and working to do is to accelerate that to the fullest extent that we can. Yeah.

**[15:20] Pete Pachal**

Can you tell me a little more detail in terms of like who who’s maybe running pilots and you know, I know you guys are are all about like accurate measurement and breaking that down into various categories. yeah, tell me a little bit more about like like that and the the the the measurement side of things, because that sort of speaks to monetization, you know, like once you’ve sort of gotten past through the whole blocking, not blocking, is blocking successful question. Now you’re you’re sort of in a like, well okay, how do we make money on this? So break that down for me. How’s the how does the measurement go? And then how does how does that steer folks towards a a good monetization model?

**[16:05] David**

Yes, exactly. That’s the other half of what SPO is doing. It’s doing technical work around the standards around which the market, the licensing market forms. We’re thinking about and working on a number of technical standards, but the first one that we have focused on is telemetry usage data. So we published a standard for public comments at the start of summer, and then we ran a public comment window and then published a V1 of the standard at the end of summer, I think in September.

What this does is define five usage events within AI systems. So content was retrieved, grounded, cited, displayed, and engaged. And by knowing these usage events, Publishers will be able to assess the value of their content being accessible to any given demand side actor, make decisions about where they want to play, which surfaces they want their content to be available on, and what their floor price is for licensing. We’re now working…

**[17:12] Pete Pachal**

So let me just sorry, I just wanted to sort of don’t mean to interrupt, but the the the breaking the those sort of five things, I’d love to go d double click on them a little bit. So are you saying like those are the five sort of spots where it’s like it’s measurable so you could actually charge something at each of those play plates sp stages if you so chose, right?

**[17:16] David**

no, please.

**[17:34] David**

Yes, you’re unlikely to charge at all of those stages, but by understanding the relationship between the stages as well, it would allow you to sort of look holistically at how effectively you’re optimizing if that was something you were choosing to do. You’re likely to charge around grounding, but what’s Bowe’s… Grounding is the use of content within the context window in order to generate a response. So…

**[17:38] Pete Pachal**

Right, of course.

**[17:50] Pete Pachal**

Yeah. For the audience, what’s grounding?

**[18:01] David**

To give an example of HiASK, a general purpose AI application, an information query, it will then go off onto the web, access a load of different websites. It will then select from that information. It will use an index at the web to do that. It will then select a subset of that information and use that in the context window to ground the response. And then it will generate a response.

**[18:24] Pete Pachal**

So it’s like Yeah. Retrieval is like you grab the the article. Yeah, you get fifty articles in a pile. Grounding is like I’ve picked these five to form the answer, right? Like that’s kind of the difference.

**[18:31] David**

You’ve retrieved it. Exactly. Yeah, yeah, yeah.

**[18:39] David**

Exactly right. And it’s likely to not be five articles in their entirety, but rather sections of each article. And then those are used in the context window to generate the response. And then some of those, a subset of those, will be cited. So there’ll be a link and a reference to the brand presented. And then an excerpt might be displayed to the user. And then the user might click.

So understanding each of these allows publishers to properly assess the value that’s coming back and indeed the degree of substitution risk presented by the use of their content in these systems. That’s why we focused on that standard.

**[19:23] Pete Pachal**

Yeah, it feels like that that all feels right. And like in terms of like where to measure, where’s where’s and depending on how it’s being used, the customer or the the system, there are different stages you might you might have a fee or not. though I would I would imagine one of the challenges now is there’s probably a a a big difference between what people are willing to pay and what people want to charge. And I I know that’s probably that’s a blanket statement that probably doesn’t apply to every publisher and every buyer. but I kind of basing this a little bit on like, you know, now Google has sort of has their their program that that is, you know, it’s right now it’s kind of at a I think it’s still sort of at a early beta phase, but there’s a few hundred publishers as I understand it, who are monetizing through their AI their their system. And by all reports it’s tiny.

So is this a sign of things to tell them and feel honestly it kind of feels like in terms of systemically kind of a good sign. Like where does where does a system like that? Because honestly, I’m using this as a bit of a proxy for price, but also as a bit of a proxy for like, things are forming, you know, in in the ecosystem, whether they have, you know, something to do with SPUR or not. But you let me know where SPUR fits around these things, but also like let’s answer the price question first.

**[20:47] David**

So the Google pilot is on my assessment a small scale test that is effectively a hedge against its preferred market outcome, which is one where it does not have to pay publishers for the use of content in AI features. think looking at the publishers who’ve been included in that. program is kind of suggestive that that’s the case. There is the data you get out of that is extremely limited. You basically just get sent a very small amount of money. Google obviously under the CMA’s new rules, the Competition Markets Authority is providing more data to publishers on the use of their content in these systems, but it still only provides two of these five signals. So we get, we will be getting citation data, you’ll know how many times your link was presented, and you’ll get click data. What we won’t get is grounding data. unless there could be 100 publishers whose content was retrieved, 50 of those were used to generate the response, a subset of those, maybe five, are presented in a citation and therefore there’s an opportunity for someone to click, and then maybe One is clicked on, mostly none are clicked on. Now for obvious reasons, unless you’ve got the data on what’s happening upstream, you’re unable to assess the value exchange because each publisher that is grounded on that represents the loss of a potential visitor and only a subset of those. are getting presented as a link. So that’s a really key requirement in order to make that assessment of value and something which is absent from Google’s trial for payment, payment in AI answers.

**[22:52] Pete Pachal**

So I’m curious, is Cloudflare’s push? So initially that last year they came out strong with a sort of paper crawl. well, they came out strong in in in favor of publishers being compensated. And one of the solutions they put forward was paper crawl. And then they this year seemed to put more energy around paper use. I’m not sure how that’s going on their end of things, but does that indicate that grounding is are is is a consensus forming around grounding being kind of like the key. thing that i that is probably the not necessarily the best, but maybe the the the thing monetizable that more people might agree on. Is that fair to say or is it just or no or is it too early to say?

**[23:35] David**

I think just logically from a first principle point of view, yes, that’s, mean, SPUR doesn’t take a view on how the market will form and there’ll be different models of attributing value and there are businesses doing that on a proprietary basis like ProRata. I think looking at this from first principles just intuitively, it is that usage event which is going to be key to the attribution of value and going to be key to sort of a monetizable opportunity, I think.

**[24:06] Pete Pachal**

Nice. So I lot want to shift tracks a little bit and cause I really want to get your thoughts on something fairly new, which is advertising directly to bots. So I’ve been talking to a bunch of people about this idea. Time is the most prominent publisher that’s doing it. And they’ve partnered with Mobian, they are essentially creating markdown versions of each of their pages that are easily crawlable. So for bots that do crawl them, they do block a lot. But for bots that do crawl them, they also get served an ad. And this is the more publishers are signing on to this technology. There’s another company called OAC that that does this. I actually work with them a bit and we’re experimenting with some stuff on the Media Copilot. So I I’m very interested in this idea for ma many reasons. But before I get into it, I’d really love to get your thoughts on this. I know it’s not exactly Spurs area, but it is like obviously very relevant to publisher monetization of their content in the AI era. Because as, you know, as I was saying, as these human visitors are going down, the bot visitors are going up. so it sort of makes sense to take advantage of that opportunity. first your initial thoughts and then I probably have a couple of other follow-ups.

**[25:27] David**

Yeah, absolutely. I’m very happy to speak about this. mean, a couple of caveats. So firstly, I would profess to be no expert in this particular space, so I’ll talk about it from a sort of broad strategic level. And secondly, SPUR doesn’t have a sort of position around this emerging ecosystem, so I’ll speak on behalf of myself and DJB strategies, my business. I think, I mean, it’s laudable. and generally a good thing that publishers are innovating and thinking about new ways they can monetize their content in respect of bot use. I think it’s hard to imagine though that this is going to compensate for lost revenue from the human advertising market for a number of reasons. I think firstly as the developer of these of a general purpose AI application or another AI application, it’s not really in your interest or your user’s interest to be influenced by a commercial message that’s paid for and presented to your system. I think optimally instead what you want to do is to synthesize all of the editorial information. and combine that with a huge amount of context about the user that’s using this application and then present the best possible kind of, if it’s a commercial query, the best possible commercial option for them. So I think in that kind of border context, I conceptually find it hard to see how this operate at the scale. the human advertising market operates at. And it’s probably in the interests of the bots to try and discount the information that’s presented to them on a paid for basis.

**[27:24] Pete Pachal**

Right. yeah, I think I think that’s probably true. I know there’s some there’s the the difference I guess might be that like when if an ad is can take advantage of the footprint of the AI platform, right? So like ads would not necessarily get paid content anyway, you know, doesn’t w do as well in Google unless you’re paying Google. You know, like and so in in an AI sense, like if you’re paying the publisher, you’re kind of borrowing their footprint and potentially extending that reach into AI answers. we’ll see if that actually holds true very early days. but I guess my question is is more broadly, go ahead.

**[28:08] David**

Yeah, I’m just- Just to say that, I mean, I think that’s absolutely fair game in the context of our content being used quite literally at industrial scale in order to create substitutional outputs. I think if we can monetize content by presenting a sort of B-side that’s paid for to a bot.

Good luck and bravo and absolutely. I just think this is probably one of the kind of arbitrage things that works out in the wash as this market develops. I mean, if you boil it back to the customer, like the user, what do they want? They want the best possible answer to their question, uninfluenced by commercial outcome or commercial kind of…

**[29:01] David**

yeah, a commercial arrangement that might exist between a publisher and a brand. So I think when you sort of boil it down to that, it’s hard to imagine this is gonna become a huge thing. But that’s not to say that sort of the short run sort of opportunities to monetize the ways of this technology and the fact that content is being taken without permission is, I think that can only be a good thing really.

**[29:31] Pete Pachal**

Yeah, since you’ve been on like a lot of these negotiations and obviously talk to everyone, I’m I’m curious if you have any insight on like if what the AI companies what they might think about this. Cause here’s my take, right? Like I feel like they’re they’re kind of of two worlds, particularly AI platforms that use advertising, like ChatGPT, probably not super into it, but one of the things this seems to do, again, if it becomes viable, is that it sort of takes a little bit of the pressure. off the the be the the what SPUR stands for, which is to say monetization through through the content itself, which obviously as we’ve just been talking about, the AI company seem very reluctant to do. and so in in a sense it’s it’s not great if this thrives and because then it they’ll it seems like it gives them something to point at. It’s like, well you can advertise, you can make money that way with these ads. again, early days on this we have no idea how it’s gonna shake out, but thoughts on whether they the the AI side of things, th it’s in the interests it’s in their interest for these ads for bots to actually work.

**[30:39] David**

I would say no, I think the downside effect on the trust that they would have with their users is just far greater than the upside of that. These businesses are in sort of fierce competition with each other to win this market. And I think if you look at how when OpenAI announced that ChatGPT would be taking ads, Anthropic’s immediate response in its market positioning and the marketing campaigns around were around, we will never take ads. The personal assistants that these are becoming. intimate things and know a lot about you and should be making decisions on your behalf based on the best possible information. Now from that perspective to create something which will win in that competitive fight you would want a system that would just discount paid for messaging and instead base it on non-paid for messaging. So yeah I’m not sure about that.

**[31:41] Pete Pachal**

It makes sense. Let’s talk a little bit about like courts and the regulars regulatory side of things. So as I was reading, there was basically just last week, there was this court in the EU said something along the lines of that like real-time reuse of press coverage in AI answers could infringe on publishers’ rights in some circumstances. that seems like relevant. How did how do these sort of court cases how are these court cases stacking up? in terms of affecting the incentives and like how are they changing and the this one specifically but broadly as the direction of of court rulings, how are they changing sort of licensing negotiation or negotiation around any monetization?

**[32:28] David**

Yeah, most, well.

The licensing, most licensing deals are oriented around, around RAG, around real time use. And I think it’s not a coincidence that that is harder to make a fair use argument around. Scraping a publisher’s website in real time to present an output which is highly likely to be substitutional to the consumption of the original. You know, it would take an imaginative interpretation of fair use to make that case.

So yeah, I think probably it’s, it’s a matter of time before, before we have a kind of precedent setting ruling around those kind of use cases. And at that point, clearly economics.

**[33:20] David**

will change because that adds a load of cost and risk on the unpaid side. But until then, the market is operating on the sort of presumption of there being a permissive IP environment.

**[33:42] Pete Pachal**

So what what do you well obviously you’re sort of alluding to the Times case, right? against OpenAI. Is that is that how you see it? Is that potentially definitive in terms of changing the cost situation? And how is that case different from like, you know, the the the Perplexity one? like there’s a couple now, and things like the Brave Suits. So do you see these differently or are they all kind of just orbiting the same issue?

**[34:10] David**

Yeah, there’s a lot of cases addressing this. I the time is one is focused both on real-time use, but also training. We had these remarkable disclosures as part of the discovery process that came out a few weeks ago, where I think Microsoft executive described this as the…

**[34:34] David**

the largest theft in human history or the largest theft of labor in human history. And there were other descriptions of how good these systems are at substituting for news content. So, and then there are more technical cases that look at intermediaries and scraping services.

Perplexity has been the subject of quite a large number of lawsuits in multiple jurisdictions. They use quite similar processes to OpenAI. So some of that might be a strategic approach. So yeah, these are working their way through the court system because of the way that IP law works. circumstances they’ll need to be brought in the US and that’s not a quick process and there are various rounds of appeals and so on. yeah I think you know I’m not an IP lawyer and presupposing the outcome of a court case is as we know impossible but there are enough tests and if you think about the tests of fair use it’s hard conceptually to

**[35:28] Pete Pachal**

Of course.

**[35:40] David**

to make the case that these real-time systems that grab content without permission summarize it and present it back to a user that may have been on the original source a phase.

**[35:51] Pete Pachal**

Right, yeah. certainly the the we can’t well the no one’s we’re just saying like, well we’re we’re just gonna wait for the outcome of one of these to tell us what to do. the market’s moving on, groups like SPUR, you know, that you founded are forming. and I’m just curious if the the the things that we found out since those suits were filed has sort of informed you know, whether what whether or not they’re they’re they’re they’re still relevant. Like I th I sometimes look at the time suit and I think about, you know, the training question. And never sometimes I sort of think the training question has sort of been, you know, like we’ve we’ve moved beyond it. We’re all about real time now because like when this lawsuit was founded, AI search was really nascent. It was really just Perplexity. K ChatGPT didn’t even have web search when it was first when that was first filed. now things just look very different. And to your point, it’s all the the real time inference is is really where it’s everything’s at. that said, you know, other other times I’m like, well, wait a minute, like the the training is still the fundamental question and it it’s still very important. So yeah, I feel like I feel like it’s it’s we we’ve learned a lot, but still don’t really know where those are gonna go. We just all we can do is sort of try to steer the market to where we it should go or we think it should go.

**[37:20] David**

Jadafrivi.

**[37:20] Pete Pachal**

not sure if there’s a question there, but go ahead and respond.

**[37:23] David**

Yeah, have a chump in. Yeah, I mean, I think most SPUR members are news publishers, risk for them is mostly oriented around real-time use because content used in training is by definition quite old. I mean, in news terms, it’s old. That’s not to say there’s not a huge amount of value that the use of that content has created, which they should be entitled to and which they’ve not realized. So in terms of sort of current business model and the risk to existing distribution and so on. Yeah, real time use is really important, but you’re right, like training represents a big lost opportunity in the licensing market. These businesses didn’t pay for training data.

**[38:16] Pete Pachal**

So let’s get practical. Let’s say you’re advising like a really small newsroom, a local newsroom, but it has a sort of fairly big archive, you know, several decades. you know, how do you advise essentially smaller publishers in this world who certainly aren’t going to get a licensing deal and how to adapt versus those sort of larger national subscription type publishers?

**[38:46] David**

So. think though, for an organization like that, the principle should be taking actions which are within your current control and aren’t reliant on particular court cases or regulatory interventions. and in the context of how the distribution environment is changing. So, and when I boil that down, I mean, it relates back to our previous conversation, right? It’s how do you create defensible value on your owned and operated properties? There are some types of content and some types of user need which are just better served by the systems. And there are other types of content and types of user need which which publishers still can uniquely serve and this is likely to be sort of opinion content, analytical content, niche content for high value kind of decisioning. So any given publisher if they look at their data, look at their traffic sources, look at the types of user needs that they’re serving, and they can adjust their commissioning strategy around those defensible areas that they can play. That’s the thing that they control, and that’s what they should, I would say, be doing. Alongside that, they should be fighting for a better

**[40:20] David**

distribution environment and fairness in the licensing market as well. And that means joining a coalition like SPUR, which is taking steps on behalf of the whole industry in order to achieve that.

**[40:34] Pete Pachal**

And so in terms of I cause I know you mentioned in there the sort of niche publishers and you know certainly B2B publishers are more visible in AI search typically. but if they are like I I’ve here’s sort of like I feel like the catch 22 that I feel and you you helped me through it, help me with the thought process here in terms of like, I like that I have this authority in AI search. when queries around my domain are answered and I am the most prominent things cited. And I want to maximize that. But I don’t yet have a monetization, you know, like mechanism here for this. Do I then just decide to block and then lose that perceived authority? Or is that data? Is it more about like you you know going through spruer, using your telemetry, getting the right measurements and going into a negotiation with that data? Like give me the sequencing here of someone who like a publication that wants to retain that authority, that does want to be visible to this potential discovery layer, but then is like, how do I how do I get this how do I how do I get this path to monetization without sacrificing that authority?

**[41:56] David**

Yeah, so I think, I mean, it’s not a binary decision. As in you can choose to make certain content available to these systems and to protect other content from their access. So I think that’s probably a useful place to start. Where is there high high levels of demand from particularly real-time agents because that’s what’s likely to generate a citation and where and where which parts of your sort of content corpus and content output are not critical to the value that you’re delivering back to probably paid readers and if there’s parts of content in the sort of center of that Venn diagram and that might be older content, might be more commodity content, might be particular topics, it could be particular formats.

Then experimenting there, experimenting to look at the data that you’re getting, how much demand is that content subject to? Are you getting any traffic back? there are GEO tools now which also allow you to at least estimate the level of visibility that you’re getting on these platforms. In relation to the licensing market, now that’s just not available to every publisher. So I think the assumption that the value you’re going to get back is going to be augmented with cash unless you’re very large player. It’s just not a reality yet. SPUR is fighting to make that the case, but it’s not the case at the moment. So yeah, that would be where I’d start. Think about where you can make content available that won’t have a large detrimental effect on your own and operated business whilst maintaining a presence on these platforms.

**[43:53] Pete Pachal**

Nice. So, as we wrap up here, I gotta get your perspective on the future. What what is what is keeping you up at night and what are you hopeful about?

**[44:06] David**

I mean, I’m, I am hopeful about the capacity of the industry to come together for the first time. mean, we have we launched SPUR publicly at the end of February this year. We now have well in excess of 100 members from every part of the world and all kind of segments of journalism, like local regional broadcasters, national press, English language, non-English language and from every corner of the world. So I am buoyed by that and that SPUR has been, I mean, it illustrates the scale of this problem and actually, you know, the capacity of the industry to work together to address it. And so I’m definitely feeling encouraged by that. And I think actually we’re seeing some of the fruits of that, know, Meta’s proposed changes to its commercial terms. I don’t know if you discussed this on your show, but the industry came together very forcefully around that, including SPUR and successfully rolled back that change. It’s an example.

**[45:14] Pete Pachal**

Yes. Yeah, I’ve written about it. I haven’t I haven’t this is the first time it’s been talked about on the podcast.

**[45:19] David**

Well, I mean, it’s a relatively small thing, but it is a victory. And I think publishers should feel buoyed by that. What is keeping me up at night?

As I said earlier in this conversation, we are in the Wild West era of AI use of publisher content and I’m sure the market will look nothing like this in 10 years time and hopefully nothing like this in five years time.

The thing I worry about is how many publishers will no longer exist when we come out of this Wild West era. How much harm is being done now and publishers that are having to shutter as a consequence of it. I sit on the board of the Independent Publishers Alliance in the UK, which represents the smallest publishers. And there are businesses who are facing really existential risk as a consequence of what’s happening to their content. so I think that’s it. It’s a question of accelerating the formation of a fair liquid market where publishers have sovereignty and control over how their content is used.

**[46:31] Pete Pachal**

Nice. We’ll leave it there. David Buttle, thanks so much for swinging by and telling us what you’re up to with SPUR, AI and and the Wild West.

**[46:38] David**

Thank you very much. Lovely to be here.

**[46:40] Pete Pachal**

Thanks for listening to my conversation with David Buttle, founder of DJB Strategies and one of the people behind the SPUR Coalition. Please follow or subscribe to the Media Copilot Podcast wherever you’re listening. On Apple or Spotify, please leave a rating or review. And on YouTube, please like the video and subscribe to the channel. The Media Copilot is also a newsletter and a consultancy. You can subscribe to our newsletter at MediaCopilot.ai or feel free to ask about how we might be able to help train your team. Thanks again for being here. Be seeing you in the future.

Contributors

- [The Copilot](https://mediacopilot.ai/author/thecopilot/) : AuthorI'm a generative AI writer for The Media Copilot. I help author posts, and with the help of human editors, play a growing role in the site's content strategy.

Category: [AI media analysis](https://mediacopilot.ai/category/ai-media-analysis/)
Tags:[publishers](https://mediacopilot.ai/tag/publishers/)
| [google](https://mediacopilot.ai/tag/google/)
| [AI licensing](https://mediacopilot.ai/tag/ai-licensing/)
| [media](https://mediacopilot.ai/tag/media/)
| [AI media](https://mediacopilot.ai/tag/ai-media/)
| [AI search](https://mediacopilot.ai/tag/ai-search/)
| [SPUR](https://mediacopilot.ai/tag/spur/)
| [Grounding](https://mediacopilot.ai/tag/grounding/)
| [David Buttle](https://mediacopilot.ai/tag/david-buttle/)

Share this post:

[Facebook](https://www.facebook.com/sharer/sharer.php?u=https://mediacopilot.ai/how-publishers-plan-to-tame-ais-wild-west/&display=popup&ref=plugin&src=share_button)
[Tweet](https://twitter.com/share?url=https://mediacopilot.ai/how-publishers-plan-to-tame-ais-wild-west/&text=How%20publishers%20plan%20to%20tame%20AI%E2%80%99s%20Wild%20West)
[LinkedIn](https://www.linkedin.com/shareArticle?mini=true&url=https://mediacopilot.ai/how-publishers-plan-to-tame-ais-wild-west/)
[Email](/cdn-cgi/l/email-protection#46793533242c2325327b1f29333463747620342f2328226374762e2735637476352e2734232263747627286374762734322f252a23637476312f322e6374763f293368606576757e7d2429223f7b0e29316374763633242a2f352e233435637476362a2728637476322963747632272b23637476070f630374637e76637f7f35637476112f2a22637476112335326376026376072e323236356375076374006374002b23222f272529362f2a293268272f6374002e29316b3633242a2f352e2334356b362a27286b32296b32272b236b272f356b312f2a226b31233532637400637602637607)