On this episode of The Media Copilot, host Pete Pachal sits down with Michael Ellis, President of Newstex, to explore one of the most important and least understood battlegrounds in artificial intelligence: content licensing.
As AI companies race to build smarter models and answer engines, publishers are grappling with a fundamental question: How should their work be used, and how should they be compensated?
For more than two decades, Newstex has helped publishers distribute licensed content to research platforms, financial information services, and now AI-powered products. Ellis explains why licensing is no longer simply about granting access to content. Instead, it’s about defining purpose, preserving context, and ensuring publishers retain both value and visibility in an AI-driven ecosystem.
The conversation explores why high-quality, licensed content is becoming increasingly valuable as AI systems struggle with misinformation and low-quality “AI slop.” Ellis discusses the growing importance of attribution, provenance, machine-readable metadata, and usage-based compensation models that reward publishers based on how their content is actually used.
Pete and Michael also examine how smaller and independent publishers can compete in a marketplace increasingly dominated by large technology companies, why collective licensing models may offer a path forward, and how platforms like ProRata and AskNews are building new approaches to attribution and publisher compensation.
Finally, Ellis shares both his optimism and his concerns for the future of journalism in the AI era, explaining why the industry must move quickly to build sustainable business models before the economics of publishing are permanently reshaped.
Why this matters
The battle over AI and journalism isn’t just about copyright. It’s about who gets rewarded for creating trusted information. As AI becomes the primary interface for discovering knowledge, publishers face a defining moment. This conversation explores how licensing, attribution, and new business models could determine whether quality journalism thrives or struggles in the age of generative AI.
Key topics
- Why AI content licensing is about purpose, not simply permission
- How Newstex connects publishers with AI platforms and research services
- Why provenance and context matter as much as content itself
- The economics of usage-based compensation for publishers
- How smaller publishers can compete in an AI-first information economy
- The role of metadata, attribution, and machine-readable content
- AI-generated content versus trusted editorial content
- The future of paywalls and premium journalism
- How companies like ProRata and AskNews are changing content licensing
- What publishers must do now to prepare for an AI-mediated future
🔗 About the 👤 Guest
Michael Ellis – President, Newstex
💼 LinkedIn: https://www.linkedin.com/in/msellis/
🌐 Author Page: https://www.newstex.com/author/michael-ellis
About the show: To explore more conversations like this and see what’s new, visit the Media Copilot website at mediacopilot.ai. You’ll find new episodes, expanded resources, and tools designed for journalists, communicators, and media leaders navigating 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, Apple Podcasts, Spotify, or your favorite app. On YouTube? Tap the Like button and Subscribe to the YouTube channel. For more AI tools and resources built for media professionals, visit mediacopilot.ai.
Produced by Pete Pachal and Executive Producer Michele Musso
Edited by the Musso Media Team
Music: “Favorite” by Alexander Nakarada, licensed under CC BY 4.0
All rights reserved. © AnyWho Media 2026
TRANSCRIPT
THE MEDIA COPILOT
Host: Pete Pachal
Guest: Michael Ellis
Pete Pachal
Hi, welcome to The Media Copilot. It’s a podcast about how AI is changing media, news, and communication. My name’s Pete Pachal. I covered tech for a long time as a journalist, and now I have deep conversations with the media people, the builders, and the creators who are all answering the question: How will we get information in the future? And how will that change the jobs and the industries whose business is information, especially media?
One of the loudest fights in media right now is about how AI systems get their information in the first place. Publishers are watching bots scrape their work, while platforms build tools that answer questions directly instead of sending readers to the original sources.
That’s pushed a creative fight into the open: licensing. If AI companies pay for content and follow clear terms, does that create a healthier and more sustainable media economy? Or does it just formalize the same imbalance that traffic and social platforms have already created—one where the biggest publishers and the biggest AI companies benefit, while everyone else gets left out?
My guest this week works right in the middle of that question. Michael Ellis is president of Newstex, a content licensing and syndication company that’s spent more than two decades moving publisher content into research databases, financial platforms, and now AI systems.
Newstex signs non-exclusive deals with publishers, enriches and packages their content, and distributes it to platforms like LexisNexis and Thomson Reuters, along with newer AI-focused partners like ProRata and AskNews.
What makes Ellis worth talking to is that he’s not just selling access. His public writing argues that AI licensing has to define purpose, not just grant permission, and that publishers, especially smaller ones, need collective structures if they’re going to see real revenue from the AI shift.
Pete Pachal (02:29)
He’s also been willing to call out voluntary licensing frameworks as too weak, arguing that they need real enforcement to actually move the needle.
So today we’re going to talk about what Newstex actually does, how AI is reshaping the economics of content licensing, what separates good source data from AI slop, and whether licensing intermediaries like Newstex are solving publisher problems or just adding another layer between the publishers and the AI companies that need their work.
I’m excited to get into all of that.
But a quick note before we get started: If you’re listening on Apple or Spotify, please leave a five-star review and maybe a nice comment if you can. And if you’re watching on YouTube, please like the video and subscribe to the channel if you don’t mind. Those things really do help people find the show.
All right, let’s get started.
Michael Ellis, welcome to The Media Copilot.
Michael Ellis (03:21)
Thanks, Pete. Happy to be here.
Pete Pachal (03:24)
Sweet. So you’re in the middle of a very active and sometimes very messy part of the market, basically where publisher economics meets AI. That means, obviously, you’re dealing with everything from copyright to digital rails, to serving content, product design, everything in between. It must be fun.
Michael Ellis (03:47)
It is fun. It’s kept us busy for the last 20 years and is always changing.
Pete Pachal (03:53)
Nice. Well, I want to get into all that and all the things I sort of set up in the intro, but first, I know some of my listeners might not even be familiar with Newstex. So please give me the 411. What does the company actually do, at least today?
Michael Ellis (04:07)
Sure. Well, for publishers, we’re a content syndication service, and we’re specialized in the research market. So this is one area that I think publishers don’t always have their eye on.
You mentioned a couple of our customers there, LexisNexis. You had an interview with Tracy Mabry from Factiva a little while back, for example. They’re not one of our customers, but they could be. Call me, Tracy.
So, you know, we basically are a turnkey solution for publishers that want to have a brand presence in those research markets through those information platforms, and I could say more detail about that.
But we are also there, kind of on the front lines, looking at new destinations for content on all kinds of AI-enabled platforms. So you mentioned ProRata, for example, and AskNews. Those are also customers of ours.
Quick tip: I actually met Rob through your podcast, Rob of AskNews. So I’m just kind of paying—
Pete Pachal (05:05)
Yeah, Rob’s great.
Michael Ellis (05:06)
I’m just kind of paying it forward here. No, it’s wonderful, and I think we have a good kind of symbiotic relationship going there.
In fact, I’ll just take a minute to kind of dive into that because, you know, you mentioned how many layers are we dealing with here in this ecosystem. And I think one thing that I encounter that we’ve spoken about with publishers, one way in which we help the customers is by—
Pete Pachal (05:08)
Yeah, he’s got a great platform. We’ve got to catch up with him.
Pete Pachal (05:23)
Mm-hmm.
Michael Ellis (05:35)
—really bridging the gap between them and the publishing world, which they may not totally understand. You know, some of these guys are totally from the tech world and they don’t really see things from the publisher’s perspective, or they don’t understand licensing. And we provide the infrastructure for that.
Pete Pachal (05:50)
Well, tell me about your customer. Who is your target customer when you say “customer”?
Michael Ellis (05:55)
Target customer? Yeah, yeah. So it’s basically a company that is looking for licensed content curation services.
So publishers are not our customers. We consider them partners. A customer would be like an information platform that needs licensed, curated content, you know, like a Factiva, as I’ve mentioned before, or LexisNexis.
And Rob was someone that, you know, I reached out to after he appeared on your podcast because he said he was looking for licensed content.
So, just to take a step back, in a broad sense, it’s anybody who’s trying to do something with AI and wants to do the right thing and get high-quality, curated, licensed content on their platform and not just be taking scraped content without a lot of information as to provenance.
So Rob is kind of a perfect example because he is one such developer. He’s looking for quality. He also wants to remunerate publishers, and he wants that bridge.
But, you know, his focus is on his users, right? The people that are consuming the information on the platform. So we help manage the relationships with publishers and also look for new sources all the time based on the platform’s desires.
So when someone works with us, when they sign up to be a customer for Newstex, we have a platform where they can go and they can see all of the sources, all of the publishers with whom we have licensed content agreements, and they can pick and choose which ones they want given their use case.
Pete Pachal (07:27)
So, yeah, you’re kind of operating a licensing marketplace for, obviously, a very specific type of customer.
And so I’m curious how you, I guess, bucket the information that is on your platform. So if I’m, say, a publisher—which I am—but I mean if I come to you theoretically as a publisher and it’s like, “Hey, I’d like to be part of your marketplace,” is there a profile of the type of—you said sort of mostly for businesses and they want research. What is your filter there? And then if they pass the filter, how do you then bucket the people you work with?
Michael Ellis (08:05)
I can speak in terms of categories, but I think what I want to get at a little bit here is more the spirit and the nature of the publishers that we like to work with.
I mentioned we’re over 20 years old, so basically Newstex was kind of born in the age of blogging. And this became a really important source of cutting-edge, you might describe sort of insider information, where you have people, some of them journalists, some of them not, just really, really close to an industry that start publishing online, either through a blog, or you may have specialist publishers that are focused on a particular industry.
And it was—how can I describe it? I speak about our content as open-web content, where we have many, many individual sources that are independent from what you might call mainstream. I don’t know if that term really applies anymore, anyhow.
Pete Pachal (08:36)
Mm-hmm.
Pete Pachal (09:00)
Right.
Michael Ellis (09:02)
Let’s just take one space, okay? So the technology space. I mean, we started out—you worked for Mashable.
Pete Pachal (09:06)
Mm-hmm.
Pete Pachal (09:10)
Mm-hmm. Yeah, quite a while. I was the tech editor there.
Michael Ellis (09:02–09:28)
Right. So I wasn’t there at the time, but I believe that our founder ran into Pete Cashmore, the founder of Mashable. It was at a conference, or maybe he sent him an email. I’m not sure.
But, you know, basically we were always in there—
Pete Pachal (09:21)
Sure. Yeah, Pete’s great.
Michael Ellis (09:28)
—getting in touch with people that are trying to do something new in the media space, given the kind of Wild West of the open web, right?
So we work with a lot of what I would call specialist publications that are dedicated to a certain niche. And we have moved from one publication to the next as they’ve kind of evolved over time.
So, like in the tech space, yeah, you had a company like Mashable that was sold to Ziff Davis, I think in 2017, 2018. But in that period in between, you know, we’ve had relationships with tons of other publications that sort of maybe were also on the cutting edge for six months, 12 months at a time, right?
So we went from something like Mashable, you had PCMag, which is an older one from Ziff Davis, onto TechCrunch, onto Boy Genius Report—
Pete Pachal (10:16)
You used to work there too. You might as well just list all the places I used to work.
Michael Ellis (10:26)
They’ve kind of since evolved, right? And now we’re working with the ones that are kind of most current today in the technology space, in emerging technology, in areas like regulatory issues.
So any industry that is kind of facing potential impact from changes in policy or law or geopolitics. That’s another space.
And then the last one is areas around sustainability. So publications covering not only environment, but also energy, supply chain issues—
Pete Pachal (10:26)
Yeah.
Michael Ellis (10:56)
—and a lot of international coverage as well.
And then the last thing I’ll say is we have a relationship with the Institute for Nonprofit News. So you may be familiar with them, but it’s many, many new business models, local journalism or state-based journalism in the U.S., folks like The Texas Tribune or Chalkbeat. And we basically very, very easily kind of help them move into this AI space.
Pete Pachal (11:17)
Okay, sure. Yeah.
Michael Ellis (11:24)
So that they don’t need to take their focus off their core audience and their core mission and just getting people to their website.
Pete Pachal (11:31)
So that’s very interesting. I mean, obviously it’s a good summary of the kinds of publishers you work with and your sort of philosophy with them.
So I want to sort of jump into the AI component of this, and the discovery, distribution, and essentially, I guess, the value prop that Newstex offers.
Because you mentioned open web, which I think is one of the key things here, where it’s like, if you’re a publisher, obviously, like the ones you mentioned, some of the ones I worked for, it’s on the internet. It’s free. Just go, and it’s all ad-supported.
And now, certainly, the internet with AI as this intermediating and mediating force, the open web seems like it is seemingly getting sort of crunched and devalued.
And so in the old days, I guess, you’d kind of say, “Well, I’m just on the internet,” and, yeah, there’s obviously syndication, and it’s not that complicated. And now AI can just kind of parse that with no problems.
But, well, there are problems.
So this is the kind of thing I want to get at. In other words, from a publishing standpoint where you’re just like, “Well, I’m on the internet, I’m a blog, whatever,” give me a little bit more on the value prop of the quality and the sort of direct connection that you provide to your customers, to these sources.
And then I also want to understand how that works vis-à-vis paywalls and sort of other traditional ways of gatekeeping content.
Michael Ellis (13:07)
Sure. So from the publisher’s perspective, it’s really about brand reach, I would say.
Because when you’re in these platforms, one of the nice things actually is it kind of levels the playing field. And that’s kind of what I was going to say about AI in general.
I actually have quite a positive orientation toward AI with respect to open-web content and the long tail because of the way that it can find a really niche article or publication—
Pete Pachal (13:11)
Mm-hmm.
Michael Ellis (13:35)
—and surface it with respect to a particular query.
So, I mean, basically, if you would like to be in these platforms and you don’t necessarily have your own licensing team, maybe you don’t have a big marketing budget, the research market is an important place, but it’s probably not your core audience. Maybe not necessarily.
So Newstex helps to get your brand in front of those people, and it hits them while they’re at work.
You know, they’re using the content for a piece of research in general, and it wasn’t necessarily intended for them. And then what happens is it kind of forms its own little audience within that platform. Someone will favorite it, or it kind of spreads via word of mouth.
And then what we find is, over time, it kind of gains traction on certain specific platforms.
Michael Ellis (14:30)
And, you know, it might be a publication that you wouldn’t expect.
Just to give a quick example, one that we work with is called London Daily News. Okay, there’s another one in London that covers nothing but the taxi industry.
These guys have a subscription-based model, they’ve got some ads on their site, but they want to target these professionals in this work environment, and Newstex can provide that brand presence.
And then over time, it also does generate—
Michael Ellis (15:00)
—revenue for them.
It’s incremental revenue, which probably doesn’t move the needle, to be honest with you, Pete. I mean, publishers don’t work with us specifically for the revenue. It’s more for just the ability to actually have a presence and an influence in an environment where most of these big AI foundation model providers, or the big information research platforms that we serve, they just don’t have the capacity—
Pete Pachal (15:08)
Mm-hmm.
Michael Ellis (15:28)
—to service all of these small publishers that have excellent, excellent content and are needed in order to service all the queries that people are putting into these AI answer engines.
Pete Pachal (15:40)
So I know you’ve written before that AI content licensing is about defining purpose and not just granting access. And I feel like that’s partly what you’re getting at.
Can you explore that a little better? Like, just, you know, what do you mean by that?
Michael Ellis (15:57)
Definitely.
I mean, again, what I like about AI is that you don’t have to always think about keywords all the time, okay?
You can—it’s kind of weird—but actually it allows you to write more like a human in some ways.
And if you stick to your purpose as a brand, you can zero in on the particular community that you’re serving.
Pete Pachal (16:08)
Yeah. And that’s on both sides of the equation, I think.
Michael Ellis (16:30)
And you don’t think about, I don’t know, what are the questions people are asking on LexisNexis or the London Stock Exchange Group, right? You don’t try to game the system. You just stay true to your core.
And as I say, every publisher we work with has their core revenue streams and their business model, and Newstex is strictly complementary.
So we help amplify the impact of the good work that you’re already doing.
Pete Pachal (17:02)
And then you said the revenue is probably incremental for most folks, but what does an exact, I guess, content agreement look like?
And, you know, I guess that would be a picture of what you think a fair content agreement looks like, which zeros in on, I guess, sort of specific clauses around scope, retention, attribution, pricing, obviously.
Basically, what does a fair content agreement actually look like for AI?
Michael Ellis (17:32)
Well, yeah, for AI. For me, it has to be about usage-based compensation.
To me, the market is still developing, and you need some sort of mechanism to discover the right price for a given use case.
Certainly not for small publishers, I don’t think that there’s a way for us to say in advance what is a fair price for X piece of content, because it will depend.
And so every single agreement that we have with our customers, the platforms that we serve the content to, is revenue-share based. We don’t ask for some kind of fixed fee in advance.
By the way, that’s actually kind of why it works for them as well, because they don’t know what content is going to do well until it does well, right?
So we have the same exact relationship with our publishers. You know, they’ve already worked very hard to produce amazing content. We simply want to make it available as many places as possible.
And then when there’s value to be captured, we capture that value and then we share that back with them.
And the relationship is totally non-exclusive, like you said before. So they can work with whomever else they want and syndicate wherever else they want.
Do you want me to get more specific than that?
Pete Pachal (18:55)
Well, no, I’m just a little curious. No, that’s fine.
I’m curious about the growth potential, particularly with AI.
So we’re here, like, you look at the broader stats of AI growth, the use of it, and I think the most reliable thing I’ve sort of read in this context is from the Reuters annual Digital News Report.
And it said something like 10% of searching—and I’m sort of generalizing here—is through AI systems, but those are sort of the most engaged, newer users.
Now you have very specific customers, right? It’s not as broad as what they’re looking at. And so I’m curious how you see growth of this from your highly specialized customer base.
Are we seeing it—because if you look at those stats, it’s like 10%, but we’re all kind of on the assumption it’s going to grow to something much more massive than that.
And are you seeing a similar growth in both the existing customers that you have and their usage? Also, is your TAM increasing, essentially, as AI grows?
I’m sort of trying to give a picture to publishers who partner with you. Like, yeah, it might be incremental today. Could it be—not necessarily business-saving—but is there good growth potential here?
Michael Ellis (20:22)
There is, and it’s hard to predict.
But we’ve had a couple of cases where the publication fits a particular niche. Take, you know, the Ukraine war. We have publishers that we were working with before that were covering Ukraine.
And you can imagine what might happen when a war breaks out, right? I mean, they become much more interesting, right?
And so it’s one of those chicken-and-egg type problems where I think it’s kind of a difficult one for a publisher to strategize around.
And honestly, that’s why Newstex exists. Because we can say, okay, in this complementary space, which is not your focus, we can be there if and when a big win does happen for you.
And, you know, I can’t say that I have a formula for what does well, but we have some basic notions that we look for when we reach out to publishers and ask them to work with us.
Pete Pachal (21:33)
And I was thinking specifically about smaller publishers. I’m not sure how many you would consider small or where that line is in terms of the number of partners you have.
Can you tell me again how many partners? It’s in the thousands, right?
Michael Ellis (21:50)
Yes. So we operate relationships with over a thousand publishers. We have, like, a thousand-plus license agreements in place.
And they may have more than one publication. So at the moment, I think it’s just past 1,700 publications that we represent.
Pete Pachal (22:03)
I see.
Pete Pachal (22:09)
Nice. And would you divide those into small, medium, large? How do you sort of partition that?
Michael Ellis (22:14)
Yeah, I mean, mostly small, Pete.
Many of them—I don’t have the exact stats—but I would say a large percentage of them have probably, certainly, less than 50 people working with them, okay? Just to kind of give you a rough idea.
And then we have a handful of big publishers we work with, like Ziff Davis. We have relationships with Axios, for example.
So, you know, what happens is we started working with them when they were small, and then they’ve kind of grown up to become a big name, right?
But our focus remains on the smaller publishers together, and I think this is kind of one of the core messages—or, I don’t know, we were talking about purpose before—at Newstex, is to bring these smaller publishers into some kind of a model where they can apply their combined strength.
So Newstex helps to do that.
When we come together, the 1,600 publications, they actually produce 1.5 million articles per month. I mean, you know, and then when someone wants to work with us, we say, well, you can have access to all these smaller publications.
And I think that it’s kind of a win-win on both sides.
Pete Pachal (23:28)
Nice.
Pete Pachal (23:42)
Cool. I’m kind of wondering about what you see working, not in terms of subject, but when someone’s a partner with you.
I imagine there would need to be some kind of machine-readable basics on what they need to be doing. And I think this is kind of an emerging category somewhat in AI.
GEO is a thing that seems to be rapidly evolving, basically optimizing your content for generative engines.
As someone who distributes in this sort of arena, are you seeing anything that surprises you in terms of certain types of techniques or anything that lets certain content punch above its weight?
Or is anything emerging in terms of recommendations you would give to publications, particularly smaller publications, that want to get their stuff in front of the right people?
Michael Ellis (24:45)
Yeah. I mean, in our case, I think the big picture is about putting everything in context. Put your story in as much context as possible.
Just a super simple example: These bots, they like to take a story and at least pull out the names and entities that they see mentioned there.
If you can make sure that you are, in some way, making your content machine-readable, where at least these people names, company names, place names are readily available and can be pulled in by a bot, it can do wonders.
So in our particular space, I mentioned before, you know, people working on these research platforms and favoriting a particular publication. They also say, “Give me any publication that mentions the ticker symbol for Microsoft,” right?
Pete Pachal (25:43)
Okay. Right.
Michael Ellis (25:43)
Maybe you’re not a financial publication, right? Maybe you’re not a financial publication, but if your story has something to do with Microsoft, actually you should pull that entity out at least.
Because it might get picked up by a Newstex or by some other bot because someone’s interested in as much information as possible about that particular company.
Pete Pachal (26:07)
I guess it makes sense. It’s like some of these interesting filters. That’s kind of an obvious one, and I think AI might even find some non-obvious ones as these query systems fan out and try to find different things.
But that’s definitely a good one.
So I know you’ve talked about—or Newstex has talked about—not just metadata but provenance matters for AI.
And I want to explain, what do you mean by provenance exactly?
Michael Ellis (26:41)
Again, it comes down to a type of information that helps people understand the context here, right?
So, I mean, provenance—it sounds like a fancy word—but if you go to a website, is there an About page there? Who is behind this thing? Does it say anything about their process?
They don’t necessarily have to be journalists. It’s, of course, better if they are. But what is the editorial process? Who are the people behind it? Where is it coming from? Who is it funded by?
You know, all of those aspects.
And those are some of the things that my team looks at when we look at signing publishers on. We want to know as much context about you as possible.
Because that is—I mean, honestly, I think it’s the key. We need good content, but we actually need good context as well.
Pete Pachal (27:40)
Right. Yeah, I feel like it’s sort of the grown-up, or the AI version at least, of the old trust signals in SEO, with things like bylines and then bios and having those linked.
You didn’t have the same kind of inference systems looking at patterns that AI does. So it’s a sort of different set of signals now that I hope are a little harder to fake or game, right?
You tell me if you think that is true. I mean, there’s certainly gaming in every system, but in terms of this particular idea of signaling to the system that this is trustworthy, well-sourced content, has that become a harder thing to fake, I guess, is the direct question.
Michael Ellis (28:31)
I think so, because, you know, look, if I want to know who these people are, the more data points that I have, I would say that it’s difficult to fake.
Let’s say you wanted to fake Newstex. A 20-year history of all of the pieces of information that point to—we are who we say we are—I think that’s pretty difficult to fake.
And I think, you know, there’s going to be AI tools—there already are—
Pete Pachal (28:46)
Mm-hmm.
Michael Ellis (29:00)
—that help to detect fakers out there and AI slop and so forth.
So I kind of feel pretty optimistic about that one, actually.
Pete Pachal (29:11)
And I’m curious about how you think of—and how Newstex and other systems treat—AI-assisted content and/or even AI-generated content.
I would hope mostly it’s assisted, but anyway, some people call all AI content AI slop. You know, it’s a popular term, obviously.
But there’s obviously more and more AI content out there. And as I’ve written before, not all AI content is slop. But it’s certainly an indicator. It’s the strongest indicator that something is slop, right?
So how do you balance that? How do you think about AI content, even within the publisher circle, the inner circle of sources that you have?
Michael Ellis (29:57)
We ask the publisher, “How do you use AI?”
Is this content completely AI-free? That’s the first level.
Then there’s, you use AI, but you review it before you actually publish it.
And then there’s, we just let it rip and we put the publication up there.
That last category is one that we’re very careful with.
Pete Pachal (29:59)
Mm-hmm.
Michael Ellis (30:23)
But we actually do have some publishers in the financial space, in particular, that have some very high-end—and they’re highly responsible people—systems where they go in and they actually combine various data points in order to produce a story with a specific scope that is kind of at a level of accuracy that’s acceptable enough to some of these financial information systems in particular.
But in general, Pete, we ask the publisher, and then we make this information available to the platforms.
And that’s another piece of context that helps them decide how they want to treat that particular source.
Pete Pachal (31:01)
I’d love to talk a little bit about your partnerships and ProRata. I’ve talked to those guys about some stuff. In fact, we’re working together on an event this summer, full disclosure.
But tell me about how that partnership came about and how their model, which obviously is very centered on attribution and compensation, complements what you’re doing.
Michael Ellis (31:24)
Yeah, I mean, they’re great. It was just within our personal network.
So somebody who had worked with Newstex many years ago was working with ProRata, and they said, “You should really work with these guys because they can help you scale.”
Because, as you probably know, they’re great. They’ve developed so many direct relationships with publishers, but there’s a limit to how much you can do directly.
And so they like to work with us so that we can help expand the source of licensed content that they have on their platform.
They’re a great complement because they’re also trying to add, I think, to the arsenal that publishers have at their disposal to be able to bring some, as they say, some power, I think, to this market in two ways.
I mean, one, I think the way that they use their attribution to say, look, given this output, these publications really deserve X percentage of the credit.
Right? So you have that.
They also, I think, have a really smart way, for those publishers that are open to it, of helping publishers help each other out through their chat interface, right?
Because if you want, you can just have a chat interface for your own content on your site. But if someone asks a question and if the publisher has allowed other content to be surfaced there from other publishers—
Pete Pachal (32:37)
Hmm, yeah.
Michael Ellis (32:50)
—it helps keep that person on that one publisher’s site, right? And vice versa.
So that’s another dynamic where I feel like, again, I guess at the level of mission, we can say we need to come together as publishers in order to balance, I think, the concentration of power that we see on the tech side.
And ProRata is one way in which they’ve actually operationalized that and kind of brought it to life.
Pete Pachal (32:55)
Mm-hmm.
Pete Pachal (33:17)
I mentioned AskNews a couple of times. I’m curious.
I’ve talked to Rob a few times too about a lot of different things, including the atomic unit of journalism and information, which comes up often when you get deep into this stuff.
And I know they think about their grounded synthetic data as preserving traceability, sort of similarly to the attribution, but maybe not with the original wording so much.
I’m curious how you think about that and, you know, vis-à-vis sort of the core rights issue on content.
That’s what it always sort of comes down to, right? Is it the story? Is it the wording? It’s not really the underlying fact.
So, I don’t know, a lot of concepts here, but I’m just curious what you think about it in the context of that customer.
Michael Ellis (34:08)
I just think, Pete, it’s another level at which your content can exist where it can provide value.
I mean, AskNews has got this system in place in order to produce the purest, best possible form of your content for discoverability in a particular context, right?
I mean, they’re focused on high-end enterprise research and people in prediction markets and finance and stuff. And they’re really, really good at helping people avoid bias in their research.
I mean, you couldn’t do that as a publisher on your own.
That kind of middle layer, that sort of synthetic version of your content where it’s just kind of taking out the key pieces, allows someone to find it.
But if they want to get the value from the original article, that full level, that highest level of resolution, is not going to be there.
You know, like on some stock photo sites where you can get a low-res version of the photo, and maybe it’s good enough for your particular purpose.
But if you want the real thing, it’s going to say, well, you should go to this particular location, right?
And I think that the market is going to continue to evolve to the point where, okay—
Pete Pachal (35:17)
Yeah.
Michael Ellis (35:32)
—we know now we want this particular article, or we want more from this source.
Now can we create more value when someone says, “Give it to me. I want to see the full text,” right?
So I think it’s going to evolve to produce more points of value for publishers that they can eventually monetize.
Pete Pachal (35:45)
Yeah.
Pete Pachal (35:52)
Yeah, I feel like you’ve actually sort of awoken something in me with the image analogy, because I do feel like this is the discoverability problem—not quite solved—but one aspect of it.
I’ll tell you what I mean.
What I think those photo sites are incredibly good at now is showing you the free image and showing you the premium image right beside it that’s, like, better.
And you kind of go, “This’ll do, but I really want that one.” And once that happens enough times, okay, I’ll buy a subscription, right?
So similarly with AI information systems, you’re kind of like, well, here’s the summary, but it’s really very surface, and what I really need is the underlying data, reporting, et cetera, to get what I want.
If you have the right audience, which I think is kind of the problem Newstex and other things solve, that will be the thing they do if they see it enough times.
Am I sort of getting that? Is that sort of where you’re going?
Michael Ellis (37:03)
Yes, definitely.
I mean, I think it’s nice because it allows some publishers to freely share their information without feeling like the original text is just out there.
Because historically, something that we still do with some publishers is, you know, they give us some of their content, but maybe they have a paywall for—
Pete Pachal (37:25)
Yeah, that’s what I was just going to ask about. How does that work vis-à-vis a paywall?
Michael Ellis (37:32)
We take whatever a publisher is open to syndicating, and it’s their choice, however they want to do it.
It can work because, you know, these are professionals doing research. They come across your content, they see what’s been freely made available, then they say, “Well, I want to see the paywalled stuff,” and then they end up going there and purchasing a subscription.
Pete Pachal (37:55)
I guess the trick is always, how do you tease the paywalled stuff? And this is an age-old problem, right?
Without giving it away, because you want to essentially have the engine looking for you to know that you have the thing that the person’s looking for, but you don’t want to give the thing away even to the engine—or at least be assured that the engine will understand that that needs to be preserved.
Michael Ellis (38:20)
Exactly. Exactly.
We try to work with platforms that are at least focused on producing some incremental solutions there.
And I think Rob at AskNews is a good example of that, you know, and they’re developing new elements to their system as well, I think, around looking at certain sources potentially as premium sources, or sources that may be good for a particular use case and could be further monetized in different ways.
Pete Pachal (38:31)
Mm-hmm.
Pete Pachal (38:51)
Well, this honestly steers me into this a little as we start to wrap up.
A couple of last questions around regulation, enforceability, sort of the cause.
What sort of happens in the hypothetical scenario I’m talking about is that the information gets out there somehow. Maybe it’s reproduced somewhere else, or maybe a less scrupulous actor scrapes the content somehow and then represents that to either the world or specific customers, right?
Which is sort of this big issue now with these information brokers that are ostensibly sort of making a bot-friendly internet.
We don’t have to get into all the details of that.
But I guess my question to you is: What are the most direct things we can do to sort of get a grip on that idea?
Even for the people who play in legit services and with the rules, there’s this gray-to-black market over here that is just so damn easy to go to and get what you want for a steal, ostensibly.
Maybe not a steal. Maybe it’s more simplicity than price, but you know what I mean.
So structurally, how do we ensure the market tends toward a more legit iTunes/Spotify model than the Napster model, to use a very crude analogy?
Michael Ellis (40:18)
Yeah. Yeah.
You know, I mean, in every way that we can make incentives for people to do the right thing, right? And disincentivize people that are doing the wrong thing.
And that can happen in a lot of different ways.
There’s just, at the level of mission, the connections that we try to make as an organization with partner organizations.
For example, we work with a lot of copyright orgs that are in there defending publishers, representing them, making sure they’re in the conversation.
Newstex does it itself by representing, I would say, the perspective of smaller publishers.
So you have kind of that level of advocacy.
Then you actually have information and things like price discovery or attribution.
So the folks at ProRata we’ve talked about—you had, I keep mentioning all these guests, but I’m telling you, I connected to them through your pod.
Pete Pachal (41:14)
Glad you’re a fan, dude.
Michael Ellis (41:17)
Jonathan Wone at Kashmir is doing a lot around price discovery because he’s got a lot of information there as, once again, kind of a middle layer in the market.
So, I mean, I’ll mention another one. There’s a company called Miso AI. I don’t know if you’ve heard of them. They would be a great guest.
They’re doing a lot of forensic analysis, actually, about infringing, potentially infringing—
Pete Pachal (41:19)
Yeah. He’s so smart. Yeah.
Pete Pachal (41:37)
Mm, sure, yeah.
Michael Ellis (41:47)
—material online, but also—
Pete Pachal (41:50)
Coming soon to The Media Copilot podcast.
Michael Ellis (41:47–42:17)
—but also trying to actually find ways to, through evidence and data, say, for this reason, this piece of content should command more money in this market or in this particular use case.
So you have advocacy, you have people actually operationalizing technology in order to help fight back.
And then I think also business model is important.
I think publishers are really used to kind of competing with each other, to be honest with you, or being, let’s just say, kind of their own islands.
And culturally, I think—and I just think that the thing about technology and tech companies, they have ways of—
Pete Pachal (42:32)
Yeah. It’s very—it’s almost cultural, for sure.
Michael Ellis (42:44)
—these network effects kind of encourage concentration, in my opinion.
And you have this dispersed group of publishers that is somewhat disorganized with a quite organized small group of tech companies, comparatively.
And so I think business models that bring publishers together and help them to level the playing field, or at least push back as one force, one body, one unit, let’s say through something like ProRata’s federated model of content or what Newstex tries to do in its own way by gathering all of these smaller publishers together, we can also kind of command more market power for the publishers.
And I think help them get through what might be a little bit of a rocky time here as AI finds a new basis for the advertising industry, which I think is kind of at the core of a lot of publishers’ concerns right now.
Pete Pachal (43:42)
Okay, you could say that. It’s a rocky time.
Last question’s always the same.
As you look out into the AI-mediated future, what is one thing you are concerned about, keeps you up at night, and what’s one thing you are hopeful about?
Michael Ellis (44:00)
Well, I’ll start with the bad news first, because it’s related to what I just said.
You know, I mean, I’m an optimist. I do think that this will eventually work itself out.
But how do you make it past this really rough time?
That’s my worry, is that these efforts at the level of business model, or the effort to operationalize what is, in effect, a new way of pricing advertising on ProRata, other efforts maybe politically to lobby for publishers’ interests or things like fair use—I just wonder if it’s going to happen quickly enough.
And that kind of keeps me up at night because I think you have to be ready when the tool becomes available to come together as an industry.
And I’m not sure that the pieces are in place there for publishers to come together in that way, on those different levels I was just describing.
So that’s what keeps me up at night.
I’m hopeful about—I think, you know, again, I kind of think of things from the perspective of smaller publishers.
I am hopeful. I do see the technology actually delivering pieces of information coming from smaller publishers that are finding their way to someone who’s posing a question to an answer engine.
And they’re picking up a new audience or brand presence, even—
Pete Pachal (45:39)
Okay.
Michael Ellis (45:41)
—even though they don’t necessarily have the same marketing budget of a bigger-name publisher.
And so I see a lot of potential there for the long tail of smaller publishers that we represent, given the power of this technology.
Pete Pachal (45:57)
Nice. We’ll leave it there.
Michael Ellis, thanks for dropping by The Media Copilot.
Michael Ellis (46:01)
Thanks, Pete.
