424: The Future of Content Creation (and Protection) in a World of AI with Paul Bannister from Raptive

Bjork Ostrom: This episode is sponsored by Clariti, that’s C-L-A-R-I-T-I dot com. And I’m going to give you a really specific example of how you can use Clariti if you sign up today. And that is poster page specific tracking of changes that you’re making. And you can use the notes area within Clariti to make a note anytime that you make a change. An example of when you’d want to do this, let’s say that you’re switching over some of your YouTube videos to be AdThrive or Mediavine video players. You want to make sure that you’re tracking to see when you look back three months later, the change or the impact that that had. And personally, what we’ve noticed as we’ve worked on content is like you forget. If you don’t have a system, if you’re not making a note of that somewhere, you’ll forget.

And so within Clariti, there’s the ability to leave a note, anytime that you’re making a change or improvement on a piece of content to allow you to go back and see how that change impacted things. There’s lots of other ways that you can use Clariti, but I thought it’d be helpful just to give a really specific example. If you want to see what those other ways are, you can go to clariti.com/food to get 50% off your first month. Again, that’s C-L-A-R-I-T-I dot come slash food, to get 50% off of your first month. You can start taking notes on the changes you’re making and explore all the other features. Thanks to Clariti for sponsoring this episode.

Emily Walker: Hey there, this is Emily from the Food Blogger Pro team and you are listening to the Food Blogger Pro podcast. Today on the podcast, Bjork is interviewing Paul Bannister from Raptive. And I know you might be thinking didn’t Bjork and Paul just talk on the podcast? And you’d be right, but they’re here for a really important time-sensitive conversation today that you won’t want to miss. Paul and Bjork are going to be talking about the future of content creation and protecting your content in a world with AI. As you know, it’s super hard to keep up with AI. The technology, the legislation, our understanding of it, everything is changing so rapidly right now and we want to keep you as up to date as possible.

As content creators, it’s really important that we stay as informed as possible to understand how AI might change the future of food blogging and content creation, and what we need to be doing to stay ahead of these changes. Bjork and Paul discuss how the current AI tools like ChatGPT and Bard are using existing content on the web, just like recipes from food blogs and how online search on Google might change in the coming years with AI, they also share actionable steps that you can be taking today to protect your content and how you might want to start thinking about the future of content creation in a world with AI. It’s really a must listen. I know you’re really going to get a lot out of it, so I’ll just let Bjork and Paul take it away.

Bjork Ostrom: Paul, welcome back to the podcast.

Paul Bannister: Hey, Bjork, great to be here.

Bjork Ostrom: This is probably the shortest period we’ve had between interviewing two guests. We’re not going to call it an emergency pod, but sometimes you’ll have these podcasts, social media posts, people get on emergency pod, emergency posts. It’s not that, but it is unique, and that it’s time sensitive to a certain degree, and the reason it’s time sensitive because as we know, technology moves very quick. As we also know, the world of AI moves very quick and one of the things that you are talking about with your team and Raptive and working towards is making sure that we, as publishers understand the implications of AI. And one of the things that you had mentioned that was so great on a call recently was this idea of Napster.

And we think back to Napster for those who are old enough to remember that, I definitely am, it was a sweet spot for me. I was like, “Oh my gosh, free music. Anytime I want it, I can just download it.” And you were saying on a call recently, this idea that the music industry really didn’t move quickly enough to react to Napster in a way where things evolved quickly, it moved past it, and the potential long-term damage that came from them not being able to move quickly, and how potentially we as publishers, creators, people who are writing content, putting it online, creating photos, putting it online, could … the story arc could kind of be the same for us potentially. Can you talk a little bit about that analogy and why we’re right now in the middle of that as publishers?

Paul Bannister: Yeah, and so, to back up even a step further, I try not to be melodramatic or sky is falling, Pollyanna, but there’s a distinct possibility that the rise of generative AI is as big of a deal for society, the world, certainly content creation and the larger industry that that is, that it’s as big of a deal as the printing press, which is sort of a ridiculous statement and why I hate to be melodramatic, but it could actually be. And I was around at the beginning of the web, which was an enormous deal, and I feel like this is as big for sure. So with that as the backdrop, I think that if you agree with the point that AI is a huge deal, if you agree with the point that AI will definitely be very disruptive, then you start asking the question, well, what are the things it’s going to disrupt and what are the challenges it’s going to create?

And I think the pure … again, in the simplest most basic sense, like content creation, which includes writing and photography and art and music and screenwriting, and it’s why the actors and everybody are on strike right now. It affects literally anything where people are creating a content of some form can be disrupted by this. So, if you agree with that as the backdrop of everything, then the Napster analogy I think is a pretty good one where it’s massive disruptive change, that only affected one particular industry, which was the music. And the music companies were like, “Oh, this is fine. Oh, we’re going to use the same old tricks we’ve done before. Oh, we’re going to sue people.” And they sued a bunch of college students and stuff for stealing music.

Bjork Ostrom: Yeah, I remember that. And as a college student, I remember being like, “Oh, no.”

Paul Bannister: Yeah. Yeah, and it clearly backfired because if you look at the … they were slow, they did the wrong things, they were reactionary, they assumed it would go away, and if you look at revenue data for the music industry, I don’t have it in front of me, but I think it took until this year or last year for the music industry to be as big as it was in 2000. They had such a massive hit because they were really behind in terms of the technology. The technology got the better of them, they reacted the wrong way. And ultimately, music is a good example. They have a valuable asset, people who create music, it’s an amazing skill, it’s an amazing art and people will pay for it if it’s packaged the right way. Finally, it’s kind of getting to a place where it is. So I think that that moment for music is now kind of multiplied because it’s for any form of content creation.

So I think that for in the scope of what we’re talking about here, people who create content online food specifically, but really any kind of online content creator, like you’ve got to be looking at AI and saying, can this make better content than me? Is this going to kill my search results? Who’s stealing my content and using it to feed these systems? There’s all these questions that are wrapped up into it that are really complicated and scary, and yeah, if we don’t actively lean into this, we all could be kind of Napterized for the next couple of years.

Bjork Ostrom: If people … Napsterized did you say, was that … yeah, I’m going to figure out how to use that word more often. I love that. The point being, if we just sit back and we’re like, “Whatever, this is just technology, it’s how technology works.” These companies, these AI companies, if there’s not any pushback against it, there’s no incentive for them to be more generous with links or crediting. And so for those who aren’t familiar, can you talk about what’s happening, when you say generative AI, number one, what does that mean? And then number two, how does a company that’s producing generative AI content get to the point where they’re able to do that?

Paul Bannister: Yeah, So generative AI is artificial intelligence technology that can generate new stuff, typically content, and the three forms of content that are mostly being produced today are text-based content, so ChatGPT is the best example, or Google Bard. You go to one of those systems you type in. One of the best scenarios I’ve been using it for is travel. I’m going to Lake George in upstate New York next week, and I typed in, “I have a family of four and I have two kids this age and this age, and we’re going to Lake George and what are some of the places to go that are in the area?” And ChatGPT and Bard gave me great recommendations of things to do, but I dug deeper on afterwards and went to some sites to look at information, but they gave me phenomenal information, so it’s generating content for me.

That’s text. You’ve got image-based generative AI, like Midjourney, people played with that or Stable Diffusion, which honestly both of them are unbelievably amazing to literally be able to type in, “Show me an image of a panda bear wearing a gangster outfit,” and 10 seconds later you’ve got a picture of a panda bear wearing a gangster outfit and it’s unbelievably good. Really cool, and then the newest form, which is still very early days though but kind of works in different forms is generative video. There’s a tool called Runway. There’s some others where you can type in very simple video things and it’ll make you a three or four second video clip out of the blue. And it’s really, really amazing.

So it’s amazing technology. I think a lot of people have used ChatGPT probably the most, but maybe Midjourney, maybe Stable Diffusion, and you can see in a lot of cases how impressive it is and how it can create things in certain scenarios that are almost as good as what a person would make.

Bjork Ostrom: Yeah, and it’s interesting because it’s helping us do what we need to do. So what we need to do as internet users is in this case, it’s information. We need information. We’re doing travel, we need recipe, we need advice. I used it for the context of a conversation, like a problem I was working through just as an experiment. I used Bard like, “Hey, I’m thinking through this. What are your thoughts on it?” And it gave me some great back and forth considerations, but the way that it does that, the reason that it’s able to do that is because it’s taken all of the content of the internet or a lot of it, most of it, the best of it. And it’s essentially just … this isn’t the technical way to describe it, but it’s essentially stuffed it into a big computer brain and then, said it’s kind of like Neo and Matrix, like learn this.

It just plugs in, it learns all of it, and then is conversational. That’s like the intelligence part where you’re able to talk to it and it’s as if it understands you. So you can use natural language, it has historical context around your conversation. Within our industry specifically, what you could do is you could go to Bard and for anybody who hasn’t done this and you’re listening to this podcast, you need to pause the podcast or keep listening to the podcast, do it right now. Go to bard.google.com or search ChatGPt and create an account. It’s free, and at least try some of these prompts just to get an idea of what’s happening. You could say, what’s a good pancake recipe or give me the best chocolate chip cookie recipe, and it will give you a recipe.

And the reason that it’s able to do that is because it’s looked at hundreds of recipes and sometimes cookbooks even that are available online and it’s compiled all of those into its own recipe. So the risk here for creators is that this gets good enough where the problem that somebody has, which is I need a recipe, the solution, that doesn’t change, they’ll still need that. The solution for how they get that could potentially change. And it could go from going to Google and searching for it or going to Pinterest, clicking on an image and instead going to Bard and saying, what’s the best recipe for blueberry pancakes? And it will give it to you. You make it and you’re on with your day, and if a thousand people do that and then 10,000 and then a million, that changes user patterns. Is that an accurate description of the risk analysis?

Paul Bannister: 100% and I think embedded in your statement are two of the fundamental challenges. One is you talked about the prompts. If you go to ChatGPT or Bard, things to try and it’s like show me a good recipe for pancakes. That’s exactly what people are doing now with search and in a world … and today, when you type it into Google, you get 10 blue links back, and if you’ve got a really good pancake recipe and you’re ranking number three, that’s awesome. You win. You get a bunch of traffic. In the world where the AI system is giving the recipe, there’s no necessity for links anymore. The recipe is just given to you there because the system built it for that person, so there’s no links anymore. So that’s issue number one.

And issue number two, as you mentioned that the system has scanned hundreds or thousands or tens of thousands of pancake recipes and is basically distilled, what is the essence of a pancake and what makes a great pancake based on 10,000 data points. So effectively, if you’ve got a great pancake recipe that was searching … ranking and search today, that recipe is an input to its ability to make a good pancake recipe, so it’s like they’re taking your content to make a pancake recipe where the system is telling you what it is and they’re no longer sending traffic to you. So both sides of the equation, you are kind of getting the short end of the stick on.

Bjork Ostrom: Yeah, you’ve helped by essentially powering the knowledge that this AI, whether it’s Bard or Bring or ChatGPT, ChatGPT being what is powering Bing, but you’ve given it that, but then you’re also literally just not given any credit. You’re not compensated for allowing that to happen, and there’s also no referral compensation for the link that’s eventually clicked, which is why SEO exists because the thought is you’re going to be able to produce content for less than what you’re getting from when somebody visits your site, whether that’d be buying a product or advertising revenue as is the case with Raptive or selling products. So the question that I have related to that, and we’d be curious to hear your opinion on is let’s use Bard as an example, being that it’s closest to Google or it’s within Google.

The way that Google makes money is by links and including links to sites. And it’s like this massive cash cow, crazy amount of money that they’re making from that. And if you use Bard right now, there’s not any of that. So it’s not like a moneymaking mechanism for it right now. My curiosity is within the context of a search engine or something that was a search engine that now is like an AI chat interface, do you think there’s any potential that Google needs to figure out how to bring links into it because that’s the only way that they’re making money and therefore is the potential with something like Bing or Google where links are being introduced and partial answers are given that result in search traffic? Have you heard any kind of industry chatter around that?

Paul Bannister: I think the scary thing … I’ll do the scary thing and then the comforting thing. The scary thing is that there’s nothing to stop Google from putting paid links into the AI results, but no organic links. And for the vast majority of the people we work with and content creators on the web in general, you don’t pay for links, you get organic links. So they could easily say, “We’re going to put our three paid links on top of the chat results, but then bury or completely disappear the organic links.” That’s like fear number one. I think on the other side, Google at least has a large vested interest in continuing to send traffic to websites. They make money on those websites too. If the results are better, users are happier, they come back more often.

There’s a lot of reasons why they don’t really want to upset the Applecart too much too quickly. So I’m optimistic that that worst case scenario won’t happen soon, but if competitive pressures increase on Google or different things start happening, they may have to move faster, but I’d be optimistic for now, it won’t be that terrible in the near future, but this is all moving very fast and hard to see when things are going to change.

Bjork Ostrom: Sure. It’s almost like if you’re using just Google as the AI engine that we’re looking at, maybe the risk isn’t there. It’s actually more for the companies that don’t have an incentive to prioritize links like ChatGPT or whatever next generative AI company comes around and trains on a bunch of content. That’s where the risk is, where it’s a completely new business model, they operate differently and people go there, they start to get into the rhythm around using that as their solution for things. Is that accurate?

Paul Bannister: I think that’s exactly right. I think that … I think Google is playing that game where they’re like, we don’t need to blow up the way search works right now. We can improve search, we can iterate on it, we can add AI based features to it. We can make it smarter, but they don’t need to blow it up right now because it’s still their cash cow and there’s a big ecosystem that they’ve built there and they don’t have that pressure. If ChatGPT figures out some amazing new search oriented chat system which ChatGPT, it’s not there yet and users start flocking to it, that’s where Google will start having to be like, “Okay, now we’ve got hard choices to make. Do we chase what ChatGPT is doing to make a user interface or do we account, how do we protect our cash cow that we have today?” So right now that pressure is not there, but it could happen at any point.

Bjork Ostrom: Yeah, what it takes is a breakthrough product that has quick adoption for users that then makes the current player, Google that’s dominating, have to react to say, we need to keep people from going over there, so we need to create something that’s similar to that.

Paul Bannister: Exactly.

Bjork Ostrom: So where does that put content creators? What do we do? What are our options? And if we talk about these paths forward, where’s the path forward where this continues to be an industry where people can build a business, be successful with it, create content and get compensated for it? Let’s talk about that path.

Paul Bannister: Right, one simple thing, I think every time I’ve been on this podcast, I’ve said the same thing over and over again, which is great quality content wins. So I think step one with all of this from a content creator perspective is keep making good content. The better your content is, the more users will like it. The more people want to send traffic to it, the more people will want to come back. There’s other forms of traffic beyond search, diversify what you’re doing, keep doing a great job, keep engaging your audience. That always has to be the most important thing for sure. So that’s like bucket one. I think bucket two then is starting to look at, again, thinking back to the examples of Napster, looking at what’s happening right now in the world.

And you’ve got all these giant media companies like the New York Times and the Associated Press and Dotdash Meredith and Hearst and others, starting to negotiate with the AI companies for different aspects of their business and trying to figure out, can we be compensated for this? Can we have certain guarantees in place? What are the things to conversations to be having today to reduce the possibility that we all get Napsterized, which I’m going to keep saying.

Bjork Ostrom: Yeah.

Paul Bannister: So I think that that’s the other part of it is like what’s the right way to engage in the larger world? To have a seat at the table and figure out what does … no one knows what the world is going to look like five years from now. We all know it’s going to look different, but we don’t know what it’s going to look like and how do we become active participants in setting that future rather than letting it happen to us.

Bjork Ostrom: Yeah, and so when you say these companies are negotiating seat at the table, I hear different companies like Reddit as an example, talking about charging for access to their site. Can you talk about what’s happening there? What does that mean when you look behind the scenes and see the deal that’s being done with an AI company? Have those deals been done or is it just negotiation right now? What does that look like?

Paul Bannister: As far as what is publicly known? Only one deal has been done, which is the Associated Press struck a deal with OpenAI, which makes ChatGPT to allow OpenAI to use a whole bunch … I mean, the Associated Press has content from the last 100 years or something like that. They’ve got enormous amounts of historical information. So they provided that data to OpenAI to use in their system, and that was kind of a paid deal. No one knows what the price was. My guess is it wasn’t that much money, but it sounds really good. That’s the only known deal. There are probably some other deals that happen behind the scenes that are a little bit less clear. We know in the image world there have definitely been deals done there, but in the text oriented world, there’s only been that one public one.

Bjork Ostrom: So in that case, what it was is OpenAI Associated Press said you can’t crawl our site. How did that … do you know functionally how that happened?

Paul Bannister: Yeah, I think part of it for the Associated Press was a bunch of the content they have, they don’t publicly release in a very usable format.

Bjork Ostrom: Got it. So they maybe … yeah, yeah.

Paul Bannister: They give them special stuff that you can’t get just by scraping a website. So they had a little bit of leverage there, where it was like, we’ve got really important valuable things and you can’t have it otherwise and it’s like, okay, that’s-

Bjork Ostrom: Behind a paywall or abstracted in some way that it couldn’t be crawled. So, it wasn’t a matter of them clawing it back and saying, you can’t have this. It was more a matter of open AI saying, how do we get this? And they’re like, you pay for it versus anything on the open web. Most of the sites that we’re talking about right now, they can just crawl that there’s nothing you can do to stop it at this point. Is that right?

Paul Bannister: There are beginning to be some things you can do to stop it. So OpenAI announced last week, or not sure when this podcast is coming up, but in early August that they have now GPTbot, which is just like this Google bot that cross sites and helps index your site for search. GPTbot is a bot that OpenAI uses to scrape your site to pull your data to use in their system, and you can now block that bot in your CMS from your site going forward. So that’s the first truly controllable thing that’s been put out in the world, but it’s very binary. You’re either like all your information is scraped or you block everything. I think part of it is I think a lot of people would say, “Well, you can take my stuff if you pay me for it.”

And again, no one knows what the model is yet for being paid, but that’s the kind of negotiation that … I think those are the early kinds of conversations that are happening, even though no one quite knows what the model is yet. For now, you can start blocking the ChatGPT bot if you want to do that.

Bjork Ostrom: And how do you do that? It’s similar to a no follow type or a no index type tag that you would do with Google if you don’t want it to index a certain page?

Paul Bannister: So it’s Robots.txt So I think you have to use Yoast or some other plugin like that to modify that on your server. I am not a CMS expert-

Bjork Ostrom: Outside of your expertise. Yeah,

Paul Bannister: Yeah, exactly. Yeah, but I think you can do it in Yoast, and I know we publish on our blog information on how to block it if you want to.

Bjork Ostrom: Great. And so Yoast, for those who aren’t familiar, it’s a plugin that is made by a company that specialized, they have lots of different plugins. Yoast was the name. We actually did a Food Blogger Pro interview with him years ago, and they have an SEO plugin. So, the SEO plugin allows you to do certain customizations and specializations. One of the things you can do for those who aren’t familiar is you can tell Google to not index something and you might think, why would you ever do that? Well, maybe you have a certain page on your site that you want to have available, like an email signup page, but you don’t want that to be indexed, but you want it to be available on your website. You might tell Google that the bot that comes and crawls your site, you don’t want that to be indexed.

You couldn’t previously do that with AI, but now, it sounds like with ChatGPT, OpenAI has released some information that is essentially like insect repellent for the spider that comes and crawls your site and you’re like, “No, I don’t want you here. Go away.” So then it just leaves, but then, the flip side of that is you just aren’t going to be included in, in this case results from ChatGPT. Is there any risk that you think with that? I mean, I think the risk would be more so with a Bard if you’re telling, “We don’t want Bard to crawl our site,” which you can’t do now and it would probably be complex, but would there be any risks to say, “Hey, we want to be excluded from ChatGPT,” and then they come out with this gangbusters feature that sends a bunch of traffic? I don’t think that’ll happen, but-

Paul Bannister: Right, right. I think the risks are things that are sort of unknown like that, maybe-

Bjork Ostrom: Yeah.

Paul Bannister: And gangbusters saying if you’re not in a system, you lose, you don’t get traffic. That’s a risk. I think if later there is some compensation model where they pay you per article for access to your content and you are out of the index, you don’t get paid, there’s stuff like that, but they’re very unknown. So, I think-

Bjork Ostrom: Yeah, it’s all speculative.

Paul Bannister: We’re recommending people do it, but it’s not a slam dunk either way. It’s definitely a complicated situation.

Bjork Ostrom: Right, that makes sense. And part of it is there would only be incentive for an OpenAI to pay for content if the majority of the sites that they want to get content from are blocking them and if not, then there’s not incentive because they can just continue to crawl it.

Paul Bannister: The pancake analogy is a great one where it’s like, I don’t … let’s look right now very quickly.

Bjork Ostrom: Yep.

Paul Bannister: Pancake recipe returns … well, this is probably wrong, but 599 million results on the web, that’s probably wrong, but let’s assume there are tens of thousands of pancake recipes on the web, even half the people block them, there’s still tens of thousands of pancake recipes that it has access to and therefore it’s hard to cut back on that access at this point.

Bjork Ostrom: Sure. Yeah, that makes sense. So there’s some things that we can do. My guess is you’ll have more control over time of what that looks like. I think the OpenAI note is a great one. I wasn’t aware of that, but we’ll be sure to link to any resources within the show notes so people can check that out. My guess is just with a simple Google search, they’d be able to figure out how to do that. The thing that’s hard for an independent creator is we’re kind of on our own. We’re just trying to figure out how we can produce content, get some stuff up on social media, get a good night sleep, and also spend time with our kids. And then you have the considerations around open AI and these paying for rights to our content. So what does that look like for the independent creator, who we can do little things.

Like you said, we can put a no index bot thing on our site to keep open AI from ingesting our content, but what are the other things that we can be doing? And I know that Raptive is kind of pushing some initiatives forward for independent creators as well. Can you talk about those two?

Paul Bannister: Yeah, sure. So the first one … again, let me step back for a second. I think ultimately, what the collective we would like to do is influence Google, OpenAI, other AI companies to come up with fair models for making sure we all keep getting good search traffic, making sure that if there’s a deal to be done around being compensated for content, that the collective, we are a part of that. And just making sure that there are fair outcomes for creators. That’s what … I think we would all say, those would be good things. We all want those things. If we’re going to influence those companies, you have to think about the ways to influence those companies. Those companies want to make a lot of money. They want to not have governments come down on them in a big way, particularly companies like Google that are enormous and have lots of government pressure on them.

They want have good press, and so they want those things. So, we think about influencing those companies through those things. How do we push them to do things that we collectively want by being in the press? How do we collectively push them to do things we want by influencing advertisers who pay them? How do we collectively push them to do things by talking to governments and saying, “Hey, this is unfair. This is what’s happening.” So those are the kinds of things that we think are important to starting thinking about influencing those companies. So step one is we release an open letter, which you can sign at protectcontentcreators.com. By doing that, you put your name on it. We would also love it for creators, bloggers to share it with their audiences.

We don’t have real people to sign it. We released it a few weeks ago. We have thousands of signatures now. We want to get to way more than that. And I think that proving to the world that people care about content creators, people want them to be compensated for what they do, people want them to be able to make a living like many of your listeners do, is really important. And we want real people to sign that because that’s a real proof point. When people believe something, it goes a long way. That’s bucket one, is the open letter

Bjork Ostrom: And to talk about that a little bit. So protectcontentcreators.com. This can be anybody, it could be my mom, it could be us, it could be anybody who wants to get behind this idea of protecting content creators in the world of AI. And what you’re doing in signing that is just saying, “I believe in this,” and you can see the specifics of what it would be. There’s kind of these three points, main points on the site. Content creators’ interest must be a central consideration, trust and transparency must be preserved, and third, content creators deserve control and compensation. So, it’s essentially this collective effort to say, “Hey, here’s this voice of thousands of people that provides a little bit of evidence, a lot of evidence towards government, towards big companies like Google to say this is important to people and they’re thinking about it. And here’s some evidence as to that.”

Paul Bannister: Yep, yep, because the reverse of that is … and that’s exactly right, the reverse of that is that the White House a month ago or so put out a whole bunch of principles for how to regulate AI. None of them talked about content creators, even though again, that’s where the data all comes from. Then, the EU recently … I don’t if the act is fully in law yet, but they have an AI act which will be the first government action around AI in the world. It doesn’t mention content creators either. So it’s like governments are already going down these paths to put regulations in place, which is probably good. I’m not saying what’s in there isn’t right, but it certainly doesn’t talk about content creators at all.

And we want to make sure that content creators have a seat at the table and real people signing the open letter is a proof point that this is a necessary component of in that case regulation, but other things as well.

Bjork Ostrom: Right, versus an individual company going and being like, “Hey, we think this is important.” It’s kind of like thousands of individuals and thousands of individuals who represent companies going to people and saying, this is important. And also a reminder of there’s thousands of these small businesses who also employ other people that this is really important to … it’s just a reminder for these people who maybe have one filter or one way to look at it that are creating … making decisions to be like, “Oh yeah, this is something that we also need to consider. Whether it be somebody at Google training a model or somebody at some government level that’s creating rules and regulations around it.

If you can tell a story of it being a team of three people in Minnesota and this is their job and livelihood, we need to consider that. That’s kind of what’s happening here.

Paul Bannister: Exactly. Job creation, economic impact of blogging and content creation online is huge and something we’re collecting a lot of data on because it is so important to get people to pay attention.

Bjork Ostrom: Got it. So you said that’s the first thing you were going to go, and talk about a second consideration.

Paul Bannister: Yeah, the second consideration, again, a little bit of backstory. About two months ago, the Washington Post had an article and released a dataset that is public information, but they basically found … they found where all the data in open AI systems come from and they write it by domain.

Bjork Ostrom: Interesting.

Paul Bannister: Two data points came out of that for us that we thought were relevant. So the first thing just to tell a bit of a story, so it’s ranked by domain where the data comes from. The biggest contributor to the OpenAI system is actually Google Patents. Apparently Google Patents at Google.com. If you care about it, has every patent ever published? It is an enormous data set. It is the single largest contributor of information to the OpenAI system. Okay, fine. The second is some UGC side, I forget the name of it. The third is the New York Times. The New York Times contributes 0.06% of all data in the OpenAI system. So think about that. Probably the biggest news organization in the world, it’s only 0.06% of the system. So it’s pretty small in the grand scheme, but it’s still the third biggest. We aggregated all 4,500 of our sites, so we’re almost 5,000 now.

And we represent collectively 0.6. So 10 times more data comes from our creators and comes from the New York Times.

Bjork Ostrom: Interesting.

Paul Bannister: So, that’s like 0.1, so we have full scale there. Second, and we’re doing this for other content verticals, but we did it for food, which is relevant to this podcast and of data in the OpenAI system that is food related, over 30% of it comes from Raptive creators. So it is an enormous, enormous, enormous component of dataset. And so what that says to us is that the collective, we are large and have power and should use that and we should really, the New York Times is out there negotiating with these companies. We need to be out there negotiating with these companies on behalf of all of our creators. And that kind of gets to the second thing, which is something we call CARA, which is the Consent to Advocate for Responsible AI. And what that is it’s a short contract that we have now given to our near 5,000 creators, and it’s all opt in.

You don’t have to sign it if you don’t want to, but if you do sign it, what it does is effectively gives us rights to negotiate on your behalf. So we can go to OpenAI, we can go to Google, we can go to the government, we can go to other different bodies and say, “We have rights from all of these creators to have a seat at this table and negotiate a good deal for them and make sure that their traffic is protected, that they get compensated, they can have control of their content, that they understand how it’s being used and all those things.” So that’s the second thing is this document called CARA, which is this short contract.

Bjork Ostrom: Yeah, and so to recap that, was the post called Inside the Secret List of Websites that make AI like ChatGPT Sound Smart?

Paul Bannister: Yep.

Bjork Ostrom: So it’s a Washington Post article. We’ll link to it. To recap what you’re saying there, essentially what they did is through some learning process, who knows how it was, they took a best guess as to where was this information being pulled from.

Paul Bannister: Is getting into super nerdy-walky stuff.

Bjork Ostrom: Sure.

Paul Bannister: OpenAI got all the data for its system from a different system called Common Crawl, and that data is all public and Washington Post got all the Common Crawl data because it’s public and they reorganized it in a way to make it more useful. So it’s not even the best guess, it is literally where the data comes from that OpenAI, ChatGPT-

Bjork Ostrom: So they went through this process to understand where ChatGPT was training their data from, and when you look at it, your point was New York Times was 0.06?

Paul Bannister: Yeah.

Bjork Ostrom: 0.06% of that entire body of content. When you roll up all of the Raptive creators, 4,500, 5,000 creators, that also was 0.06 or 0.6?

Paul Bannister: 0.6, so 10 times.

Bjork Ostrom: Okay, so 10 times more, which is just incredible kind of the power of the collective voice of 4,500, 5,000 sites. When you look at just the food vertical, which we’re talking about, it represented 30%. So suddenly, it’s like this massive amount of representation for just this Raptive creators. And you have that data because you have all those sites, but my guess is if you expand that out to just independent creators that it becomes 50, 60, 70%, probably more of what ChatGPT is training on is independent creators’ content.

Paul Bannister: Exactly. Exactly.

Bjork Ostrom: So a lot of the value from anything food related is coming from what we’re creating individually. It’s all wrapped up. It’s like put into this model which then spits out a result based on whatever the prompt is that people have to ask. You were saying creating this initiative, CARA, C-A-R-A to speak on behalf of, in this case, Raptive creators, any of these situations where there is the potential to say, “Hey, we need to get compensated for this, or whatever it might be, TBD on what that is.” How about for the people who aren’t able to do that, that being a Raptive exclusive thing, you’re just an independent creator. Maybe it’s like the Napster comparable, is like you don’t have a label and you’re not signed onto something.

Are there other options for what you can be doing or looking at? What should you be considering? Should you look your attorney up in the White Pages?

Paul Bannister: Yeah. It’s a good question that I don’t have a great answer to yet. I mean, this is all moving very fast and it’s very much at the beginning of all of these things right now. As we go out and start negotiating in bigger ways, who knows, maybe we negotiate things that are broader or OpenAI comes with the system that applies more across the board. None of us quite know how it’s going to shake out. So I think for now, I think probably the most important thing is just be educated, learn about what’s going on, pay attention. Obviously, our blog and other sources have good information about things as they’re happening. So I think that’s probably the most important thing. Then, as this all evolves and changes, there may be other options for people in the future too.

Bjork Ostrom: Yeah. How about, has there been any noticeable impact from a search perspective that you’ve seen from the data that you have? It still is that … in my world, at least, when I’m talking to people, I know it’s probably not in your world, but as soon as we get outside of the context of digital publishing in our blog and stuff, people are like, “Oh yeah, AI, I’ve heard of it but how does it work? What do you do with it?” Any of my friends are like, “Yeah, I’ve heard of it.” And I was talking to my friend who’s a professor and I was like, probably kids are using it, and he sent me this block of text and he’s like, do you think this is AI generated? And I did this AI generative test and he is like, “98% probability that this is.”

So you’re starting to see it, but not necessarily … it’s not ubiquitous. It’s not search where everybody is doing it every day, but have you noticed any data around that seeping into search or not yet?

Paul Bannister: Not yet. There is … Google has an experiment running and they call it a search generative experience, SGE, that there’s a beta test you can join and you can check it out if you want to, and it’s sort of Bard but on steroids, but really is a much deeper version of modifying the way search works. If that SGE system was deployed to search, it would definitely be very bad for creators of all shapes and sizes. It is far fewer lengths. It is far more the answers on the page, it’s vertical specific. I think food actually is a more protected content vertical right now, like bots don’t have taste buds so … well, it has a lot of pancake recipes. It really doesn’t know what makes a good pancake at the end of the day. It’s getting better and better and better.

So if that SGE system was … became the main search, it would be very bad. Again, I’m sort of confident that Google isn’t going to upset the Applecart that quickly here. So I think what they’re going to do is take components of that and point into main search over time, but it’s going to take time. Some people are worried that that’s going to happen as soon as the end of this year. Again, we’re more optimistic that it’ll take longer, it’ll be smaller, it’ll change more bits and pieces. Google is certainly thinking about search experiences that are way deprioritizing content creator results.

Bjork Ostrom: Yeah, because they … and you see this in different instances. They’re incentivized to provide the best answers possible while also getting people to click on paid ads.

Paul Bannister: Exactly.

Bjork Ostrom: Generally speaking, that’s what Google is trying to do, and if they can find ways to provide an experience that will keep people using Google while also continuing to click on ads, that’s what they’re going to optimize around.

Paul Bannister: Exactly. Exactly.

Bjork Ostrom: Also, the interesting thing within it is also need to figure out how to continue to incentivize creators to create content. In the future, five to 10 years from now, if people aren’t incentivized to create content, then these engines aren’t going to be able to be trained. Not that we know how this will all play out, but it’ll be just so interesting to see like-

Paul Bannister: It’s very right.

Bjork Ostrom: How do incentives change? How does alignment change? And it’s part of the reason why … to go back to the Napster analogy, it’s important to have a voice right now because those decisions are being made and the trajectory is being set, directionally speaking as to where that goes, does that feel accurate?

Paul Bannister: 100%, 100%. I think, yeah, no one knows where this is going to end, but it is going to be a big set of changes and we think the most important thing is just to be … have a seat at the table and be a part of the conversation and not just sitting back and hoping for the best.

Bjork Ostrom: Yep. That’s great. So as we wrap up, I think it would be good to have an outline. What does it look like for us as creators to continue to do our job well, like you said, it’s like this could be a very big thing, it is a very big thing. It just is. It’s significant. It doesn’t mean that everything is changing right now, but things are changing. So, I think what I’m trying to get at is how do we get to a place where we can stay the course, but also be aware and be educated and maybe recapping what some of those things might be. So for instance, you had mentioned continue to create good content. At its core, whether it’s written on a site, on a blog, on YouTube, the business of engaging with people and being human and connecting with people, that will continue to exist even though what it looks like will change.

Paul Bannister: Yeah.

Bjork Ostrom: So figure out how to continue to get good at that, creating good content, content that connects.

Paul Bannister: I would also have one tweak to that a little bit that I think is more and more important, and the word human is so important here. The world doesn’t need the 47821st pancake recipe. We’ve got it cover. We’ve got our pancakes, but if you as a person are special and interesting and other people like you and believe in you, your pancake recipe can be better because those people care about you. So I think it’s also like, it’s make great content, but be you and be unique and find your audience that loves you and make that new next pancake recipe great and special, not just number 47,821.

Bjork Ostrom: Yeah, it’s the difference … Content continually in and of itself, information becoming commoditized.

Paul Bannister: Yeah.

Bjork Ostrom: It’s available, people have it. It’s hard to create an incrementally better pancake recipe, but you have a strategic advantage in being unique with who you are, but the interesting thing with that is it’s a little bit of balancing of these two worlds. One of the worlds is search optimization, which is looking, what are other people doing? How do I rank higher? How do I include that, which is important and it works, but also within that, to not get sucked so deep into that world that you create this faceless content that nobody connects with, even if it ranks well, so there’s this yin and yang balancing act that exists, which I don’t have an answer for it, but it’s just acknowledgement of it.

Paul Bannister: It’s a real challenge, but I think if you could walk that line, that’s where the biggest success is.

Bjork Ostrom: Yeah, and as much as possible figure out how do you position the content that you are creating in a way that is unique, not necessarily just because of the content itself, but also because of who you are. Maybe it’s like your approach to life, how you connect with people, the angle that you take. I talk about this all the time on the podcast, but Kevin Kelly always talks about 1000 True Fans. How do you get those people who are like, “I love this site, I love this person. I love following this because of whatever it is.” How they write. They’re really funny. They always make things in this certain way, like having an angle and being human. I think that’s great, but also, within that, staying educated. So that would be another piece like creating great content, being human, but also not putting your head in the sand and pretending like it doesn’t exist.

You mentioned Raptive’s blog as a place to go. Where are the other places that you can stay educated as to the things that are happening in this world? Twitter probably is a great place for these conversations that are happening.

Paul Bannister: I’ll give a shout. I think Search Engine Land, which is a site that covers news about search engines, particularly Google. If You want the slightly deeper, more technical stuff, it’s a really good site and has tons of good information there on what Google in particular is doing around AI. Then, I think just Google searching in general for what’s the news around ChatGPT, around Google Bard and just following those things.

Bjork Ostrom: Yeah, that’s great. So I see even here the most recent or a post that came out today while we’re recording Google search generative experience, which you were talking about, adds SGE while browsing. So like, okay, there’s this new update on SGE interface that you talked about and they have some more information on that, so that’s great. So shout out to Search Engine Land. Anything else that you would add for us to be considering for those in the content creation world? The two pillars for this specifically, “Hey, stay the course. Think about creating human centered content, connecting with an audience, not just blindly going after search results to go after search results, but also thinking about how are you connected.

And then how can you be kind of the human competitive advantage, follow along, stay educated, Search Engine Land. We will continue to talk about on this podcast WrapIt blog. Google has their own blog where they talk about stuff that they’re releasing. Any other considerations for us as creators as we go out into the world in regards to this category of content?

Paul Bannister: One of the thing that’s connected to being human I think is also diversifying traffic, which we talked about a little bit before. Search is huge. That’s not going to go away. You got to make content for a search and watch that line that you were talking about before, but I think the more you are you and the more you are unique, the more that people care about you. Then, you can get some traffic from social and you can build your email list and what are ways that you can get traffic? That search is going to be the bulk of traffic for people, for the time being for sure, but every couple of percent that you get from somewhere else is really good and something you can own more. And again, I think uniqueness and being you, I think is a way to help with that. So it’s just a side point, but I think it’s important too.

Bjork Ostrom: Yeah, and the fun thing to think about there is the game you play is different. I don’t think about SEO for the podcast and I probably should to some degree. There’s probably considerations around search optimization for podcasts, not within Google, but maybe within a podcast player or things like that, which we do have descriptions and things like that. The podcast is a very different game that doesn’t require me ever to do deep search optimization. Same could be true for the social platforms. They all allow you to play a different game, and so the only game is in search. And I think it’s an important one, especially for sites where you’re publishing stuff, but it’s not the only game. So I think that’s a good reminder to think about diversification on different platforms, not only from a traffic standpoint, but also just from a fan and following standpoint.

Paul Bannister: Yes, totally.

Bjork Ostrom: Paul, if people want to follow along with you, I know that you’re on X. Do you call it X yet? Have you transitioned over into calling it X?

Paul Bannister: I do call it Twitter until the day I die, I think.

Bjork Ostrom: Okay, but other places that folks can follow along with you or Raptive or the team?

Paul Bannister: Yeah. I think for me, Twitter is the best place to follow along with what I’m talking about, pbannist is my handle. And I know for Raptive, I know our Instagram is very active, and Twitter, we post things on LinkedIn, we post things on, so definitely check all those out.

Bjork Ostrom: Cool. Paul, thanks so much for coming back on. Really appreciate it. Great conversation. That was good to nerd out.

Paul Bannister: Always good to nerd out.

Bjork Ostrom: And we’ll be in touch again soon. Thanks.

Paul Bannister: Thanks very much.

Alexa Peduzzi: Hey there. Alexa here, and thanks for tuning into this episode of the Food Blogger Pro podcast. We really appreciate you being here. And if you really liked this episode, we would so appreciate you leaving us a review on Apple Podcasts. It helps the show, get in front of new listeners and it just makes us really happy. We read each and every one and it’s just so great to hear from you what you’re liking and what you would like us to improve or change in upcoming episodes. So all you have to do is go and find the Food Blogger Pro podcast on your Apple Podcast app. Scroll down to the readings and reviews section, and then you can rate the show and then, leave a written review if you want to be even more awesome. And while you’re there, we would really appreciate if you subscribe to the podcast so that you never miss an episode.

Maybe you talk about one of your favorite interviews on the show, or maybe you just talk about the show as a whole, but regardless of what you talk about in your review, we appreciate it so, so much. So thanks again for tuning in today. We’ll see you next time and until then, make it a great week.