Jonathan Gillham is the Founder/CEO of Originality.ai – they provide a complete toolset that helps Website Owners, Content Marketers, Writers and Publishers hit Publish with Integrity in the world of Generative AI. They also specialize in AI detection technology. He is also the co-founder of Motion Invest where they help you buy and sell content sites quickly.
Jon has been involved in the SEO and Content Marketing world for over a decade. His career started with a portfolio of content sites and he recently sold 2 content marketing agencies. Through these experiences, he understands what web publishers need when it comes to verifying content.
Having been one of the earliest adopters of generative AI through his agency, Jon understood the wave of plagiarism problems that was coming, even before Chat GPT and GPT-4 were released.
What you will learn
- How AI detection tools can identify AI-generated content in search results, with current estimates showing about 14% of Google search results are AI-generated.
- Why the rise of AI content creation is changing plagiarism detection and management, as AI-generated content becomes more prevalent than direct copying.
- In what ways AI tools like ChatGPT are increasingly being used for research and information gathering, potentially challenging traditional search engines like Google.
- What the ongoing debates are about the ethical use of AI-generated content, especially for sensitive topics like product reviews that could impact health and safety.
- How AI content creation tools are forcing companies to develop new strategies for content optimization, fact-checking, and maintaining editorial standards in an AI-dominated landscape.
Transcript
Jeff Bullas
00:00:08 – 00:00:29
Hi, everyone and welcome to the Jeff Bullas show. Today I have with me, Jon Gillam. Jon is the Founder / CEO of Originality.ai, they provide a complete toolset that helps Website Owners, Content Marketers, Writers and Publishers hit Publish with Integrity in the world of Generative AI as well as specializing in AI detection technology. He is also the Co-Founder of Motion Invest where they help you buy and sell content sites quickly.
Jeff Bullas
00:00:30 – 00:00:59
Jon has been involved in the SEO and Content Marketing world for over a decade. His career started with a portfolio of content sites, recently he sold 2 content marketing agencies. Through these experiences he understands what web publishers need when it comes to verifying content.
Jeff Bullas
00:01:00 – 00:01:35
Having been one of the earliest adopters of generative AI through his agency, Jon understood the wave of plagiarism problems that was coming, even before Chat GPT and GPT-4 were released.
Jon, welcome to the show. You’re right. In the, in the hot spot at the moment with what’s happening in the industry digitally. So Jon, how did you get into this content marketing thing? And then the A I thought where, where was this sort of um where did this journey start? Because uh you were a mechanical engineer by university training and education. So uh tell us about that.
Jon Gillham
00:01:36 – 00:02:12
Yeah. No, thanks Jeff, thanks for having me. Um Yeah, so the journey was um yeah, mechanical engineering out of school, worked in uh in the oil and gas industry. Um knew that I always wanted to get back to my hometown. And so it was like many people sort of looking at some, you know, at the time that that felt like side hustles um sort of looking at uh learning new skills that could be applied online. And uh what I sort of stumbled into was um creating content for websites that would rank in Google and then generate, generate money off the back of that. Um And that was sort of the,
Jon Gillham
00:02:12 – 00:03:02
the starting thread that has been sort of pulling on, pulling on ever since. Um So we had a portfolio of sites and then built up a content marketing business where we’re buying, where we’re purchasing content and working with writers and then selling that content to, to clients and then um build up a marketplace commotion, invest where we buy and sell websites and act as a broker for people. And then uh yeah, the latest effort um is, is around originally um dot A I which is about um providing transparency between the, the writer, editor, publisher and ultimately the the the the client um to make sure that that sort of quality content is, is uh fully uh fully understood from, from start to finish.
Jeff Bullas
00:03:02 – 00:03:44
So, I mean, I’m curious about how you said, OK, I want to start a side hustle. So uh cos that’s an interesting part, especially in a digital world, creating a digital side hustle where you can actually do it online and take it with you wherever you go. It’s never been easier to do that sort of side hustle. So, yeah. So how did you choose to get into the content side? What was, was there a book you read? Was it a friend you talked to uh how did you decide to do that as in create content and create content sites and then look at how I could monetize it. Where did the inspiration come from for that?
Jon Gillham
00:03:45 – 00:04:28
Yes, this was back to like 2007, 2008 timing. So this is going way, way back. Um I was working at a hard tech start up. So it’s not, not the kind of start up that you would uh like the kind of start up you’d see a mechanical engineer in. Um And I was working as a co op in that hard tech start up. And it was like, we need a website, know nothing about that. But I was the intern that was the least important to be working on the core product. So, uh, that’s, that’s what I started working on and, uh, through that process, learned about, uh, Odesk, which is now upwork and the ability to hire sort of talent from around the world. And then, um, great, built a
Jon Gillham
00:04:28 – 00:05:06
probably a terrible website in today’s standards, but built up an ok site in that day’s standards. And then ok, we need some, some traffic and instead of a four month co-op we were able to get a site, get traffic and get a customer. Um and that from all sort of remote with no skills around from around the world. And that uh just really kind of opened up that curiosity window that uh it was like, wow, this, this, this process of, you know, distributed talent um applied to a problem and, and getting customers from, from Google uh was, was pretty fascinating. And then, um you know, I was
Jon Gillham
00:05:07 – 00:05:29
like, many students are broke, pretty broke students um outside after school and, and you know, the requirements for my entrepreneurial endeavors required a near near zero budget. So those were given that training and those that constraint. Um This was the, this sort of seo content marketing was the right fit for me.
Jeff Bullas
00:05:30 – 00:05:40
Cool. So you remember your first customer. So um and what, and that customer paid for, what consulting, what was the first customer? And, how did you generate revenue?
Jon Gillham
00:05:41 – 00:06:16
Yeah, so that, in that four month co-op where I sort of where I first sort of discovered this world of building websites and getting traffic from Google. The customer was for the hard tech company. So it was a laser scanning solution, an underwater measurement solution that that customer came in and purchased. Um After that, I was heavily working on my own, my own sites trying to drive revenue through display ads. So we weren’t selling consulting or, or anything like that. But that sort of first sale came through for, for that um that hard tech company that I was a call at.
Jeff Bullas
00:06:17 – 00:06:22
Ok. Very cool. So, in other words, you discover the power of content to attract a customer.
Jon Gillham
00:06:22 – 00:06:23
Yes,
Jeff Bullas
00:06:23 – 00:07:00
exactly. Yeah. And that was all exciting back then because I started my blog, Jeff force.com uh back in 2009. And uh and your timing would have been good then because I experienced the same sort of, I suppose, synergy and opportunity out of that. And I started writing about social media and uh i it was just a passion project but yours was doing it for someone going, hey, this works and then with a thought process, then if this sort of process works for my employee employer that I’m doing it for, it might work for me. Was that the thought process?
Jon Gillham
00:07:00 – 00:07:30
Exactly. Um, yeah, that, that was exactly. It is. And then, so going down the rabbit hole? Ok. Well, I like all these websites that we’re visiting. How do they make money? I now see this process on how to hire. Hire a website developer. Hi. I could do a little bit of it. Hire a writer, um, and get traffic from Google And now it was just, I wanted to do. I didn’t have a hard tech startup to sell a product from. But so how do I monetize it, the traffic that I could get?
Jeff Bullas
00:07:31 – 00:08:09
So, so you’ve moved into your own site and as we know that back then, um, and today you can actually buy a free wordpress template. You can, you know, learn and wrangle your way to actually build a website, which I built my first as well with a little bit of help from a friend. And I think it cost me, I think the domain name and that was about it. I think it was $45 at the end of the day for uh that. So again, I didn’t have a lot of money back in 2009. So you start these sites. So you had a hard customer buying a product? Ok. So how did you make money out of your sites when you started building them?
Jon Gillham
00:08:10 – 00:08:52
Yeah. So, uh display ads, um, was, was some and then uh the one that I had some sort of the first signs of sort of some modest success and not, not ready to leave my day job yet. But um was uh uh spreadsheets. Um So downloadable Excel templates spreadsheets that perform some calculations that were needed for um mortgage applications and mortgage refinancing those spreadsheets that sort of hey, input all this information and it’s going to spit up the answers that you need to put into on, on this form. That doesn’t provide the template for you. Um That, that was where um
Jon Gillham
00:08:52 – 00:09:08
again and monetizing through, through display ads and making those spreadsheets available for free and getting put back as a result. Um you know, pretty common strategy now, but that was uh that was sort of the first learning of sort of that, you know, engineering uh for, for uh for seo,
Jeff Bullas
00:09:08 – 00:09:17
right? Cool. So did you build the first spreadsheet? In other words, you gave ways like simple spreadsheets for free initially to attract, I suppose, build an email list.
Jon Gillham
00:09:18 – 00:09:46
Uh Yeah, I didn’t focus much on the email list. Um had one but that wasn’t the majority of the focus I was trying because I did have a uh time intensive day job by that point that also had some fluctuations where I would have like ridiculous work hour weeks. I was trying to sort of keep it as um as simple as possible from uh from sort of a monetization standpoint and something I could step away from. And so it was primarily monetized through, through display ads.
Jeff Bullas
00:09:46 – 00:10:08
Right. Ok. Right. So that was your first, so your first bit of revenue came in? You went, yay. This is working even though it’s smaller, I think I remember sold my first ebook or something and I went or an affiliate revenue or something and I went, it was like $10 I think II I think it was an Amazon check actually that got sent to me from, from the USA to Sydney, Australia. Um
Jeff Bullas
00:10:09 – 00:10:52
um So, but that’s it, I suppose it’s that first little, hey, I can make a bit of money out of this. Going. Well, if I can be better at this, then I can make more money. Um But this was also the time and I’m curious about this, this is the time when social media was starting to take off and it was very organic for about the next five years. In other words, you could do a tweet or whatever and actually people would, but, you know, actually see it before Google put in its or Facebook, put in its algorithms to actually promote its own ads and make your ads just or your, you know, tweets or posts disappear almost. So, did you use social media at all? Were you leaning into social media?
Jon Gillham
00:10:53 – 00:11:35
I think I dabbled. Um But no, I, I think, you know, any any attempts were so lacking, like probably not my bad in terms of lackluster attempts that didn’t produce the kind of results that I was seeing from an organic content approach. Um And then it felt like the sort of the social wins were and I mean, Google is incredibly, incredibly uh volatile as well, but it felt like the social sort of algorithms were, were very volatile and what strategy was working one day would struggle the next. Um It felt far more sort of um
Jon Gillham
00:11:36 – 00:11:57
less compounding than, than getting organic content in Google right now. I think I was a little bit wrong in that in terms of how long, how long there’s not many posts from 2012 that are currently ranking in 2024. So I think that that’s not necessarily true, but I think it’s at least more compounding than, than social was, is what was my sense at the time?
Jeff Bullas
00:11:57 – 00:12:39
Right. Well, I, I think back then it was actually you’re able to get Google authority a lot easier than it is today because it was also less crowded. So you’d very much doubled down on content as a seo strategy. Is that right? Right. Yeah, it’s um I remember I used hubs, hubspot had a great website back then and I almost lived on that platform just trying to seem to improve my grade by doing certain things. And uh it was a free tool that I went, wow, if I do this and I get more traffic and then I went from 10 visits a day or one or 2 to 10 to 20 to 30 to 40 I think at 50 I went, you know, almost, I cracked a bottle of champagne. But
Jeff Bullas
00:12:40 – 00:12:46
The reality was that it is a very long game, isn’t it in terms of earning authority with Google?
Jon Gillham
00:12:47 – 00:13:29
For sure? I mean, I think it’s the, the, the time horizon that you need to be having with Sco and, and Google is, is very, very long term and more so and continually more so more focused on the customer and their problems in the audience than ever before. And I think that’s been a trend that has continued. I think that if I were to think about my focus around content when I started, it was like 75% Google, 25% the visitor. And I think now it’s like 90-95%. Is this gonna be good for the visitor? 5%? 10%? Making sure we’re doing the right stuff in the eyes of Google.
Jeff Bullas
00:13:29 – 00:14:07
Yeah. Yeah, you gotta play the game, which is a long game. Number one and number two is playing the Google game because they can penalize you just like Facebook can remove you from its platform if they think you’ve misbehaved, which sometimes is a dark art because you said what happened there. What did I do? So the trouble is though the algorithms, the machines are starting to run the show, not humans. So this is the challenge. We’ll get to a I in a minute and, and discuss that. But so you obviously are learning some skill. I said I’m, I’m good enough now to actually start a content marketing agency. Is that how it is? What happened next?
Jon Gillham
00:14:08 – 00:14:48
Yeah, that was exactly. I mean, we scaled um started building a team um in, in, in my engineering role at the time I was working in uh project management um running some what seemed like crazy dollar value projects like, you know, 10 million $50,000,100 million spends um and, and running sort of AAA involved in running of a team that was, that was sort of spending that amount of money. And I’m like, OK, well, if we can apply that same project management skills set to this other skill set that I’ve been developing on the side, can I scale out um my, both my portfolio and then build extra capacity into my content writing team
Jon Gillham
00:14:48 – 00:15:09
and then sell that extra capacity and use all the systems that we were building up and, and sell that with producing content that would predictably rank well in Google using all the, all the modern tools of the day at that time. And so that’s, that’s what started the, the content marketing agency was just that sort of extra capacity from my own team,
Jeff Bullas
00:15:10 – 00:15:29
right? Ok. So, um you built that up and uh you would have learned a lot along the way, wouldn’t you by doing that with, on your own? But then you decided I wanna get back. Well, were you back in your hometown by this time? Is that so,
Jon Gillham
00:15:29 – 00:15:54
uh 2017, so I started sort of the journey in 2018, 20 maybe it was 25th. I should remember. It’s 2015, 2016 was, was the move. And so it took sort of about 55 to 7 years to kind of get, get the income up to sort of income and runway up to the point where it was like, ok, time to leave that career and, and pursue, pursue online stuff full time.
Jeff Bullas
00:15:54 – 00:15:59
So let’s fast. So you sold that agency, didn’t you?
Jon Gillham
00:16:00 – 00:16:04
Yeah. So, the agency in 2017 sold it in 2021.
Jeff Bullas
00:16:04 – 00:16:13
Ok. Cool. A and was that uh and was that just sold remotely, were you in, you know, Toronto? Were you, where were you?
Jon Gillham
00:16:14 – 00:16:20
Yeah. So it’s uh it was sold to a Canadian company but they, they were based uh out in, out in Vancouver,
Jeff Bullas
00:16:21 – 00:16:44
right? Ok. On the other coast a long way away. But that’s the beauty about um digital, a digital business is that you can run it from anywhere in the world. And quite often you’ve got remote workers that, you know, could be in Hungary, Philippines, India. It really doesn’t matter. So, the question I do have too is did you do any of the writing yourself or? Did you just outsource all of it basically?
Jon Gillham
00:16:45 – 00:17:21
Uh, I’m an absolutely terrible writer. Um, it would be pretty shocking if my sort of high school English teacher knew the amount of, sort of like, content that I was involved in the creation of. Um, it would be, they, they would, they’d be shocked. Yeah. I’m a, I’m a terrible writer so I do write some of my own stuff. Um, but yes, no, not. I, I don’t, I don’t particularly enjoy it. I don’t mind, sort of putting down trying to, like, trying to simplify complex thoughts is, is interesting and, and challenging and fun. Um, but the actual words do not flow the way that they, they, they’re supposed to.
Jeff Bullas
00:17:21 – 00:18:04
Ok. Right. That’s absolutely fine. We’ve gotta find our strengths. It’s, um, we’re told sometimes in corporations we’ve gotta be good at everything or we’ve gotta be able to do, you know, do a spreadsheet as well as do marketing. Go hang on. I can’t be an accountant and a marketer at the same time. It’s not gonna work. So, like, it’s like my brother’s an accountant and I’m a bit of a marketer writer. And, um, yeah, we’ve got two different headsets basically as in different heads, different ways of viewing the world. Ok. So you’ve sold the content marketing agency. Um Had you already had the idea of the plagiarism originality. A I then when, when did you sort of start that journey?
Jon Gillham
00:18:05 – 00:18:38
Yeah. So, that journey started. Um I, I don’t know what if you used copy scape, the plagiarism checker uh sort of the pretty, pretty dominant plagiarism checker in the world of, of what publishing, but really hasn’t been updated since what it launched in kind of 2004, 2005. It was a frustrating tool to use within the, the, the agency. And so that was one of the starting points of, of our, of the, the like, you know what we’re going, what was again, and this was not going to be the core focus originally intentionally or originally,
Jon Gillham
00:18:38 – 00:19:14
it wasn’t going to be the core focus uh for my own time. Um But building a modern plagiarism checker um that would work well with a publishing content publishing team that was bigger than one or two people, that was sort of like I knew that need was there and we started working towards that. And then we also had been one of the, at the, at the time, the heaviest users of jasper within that content marketing agency. We had an A I writing service where a sort of a transparent A I writing service and then said, oh, can we add in
Jon Gillham
00:19:14 – 00:19:52
this additional Q A QC step of, is this content A I or human? Is it plagiarized or not plagiarized? And then that’s sort of the journey that originality has continued on is taking any step that a copy editor would do as a final Q & A QC step for a piece of content and adding in tooling to be able to do that. But so that, that was sort of the um initial plan, was A I detection plus plagiarism checking and sort of a modern, a modern plagiarism checker. Um was the, was the um intent and then, and then JCPT launched.
Jeff Bullas
00:19:53 – 00:20:09
So how long, how long were you running the plagiarism checker uh sort of services focused business. How long before Chat G BT came out? Had you, were you doing that as a competitor to the clunky old legacy system of Copyscape? How long were you with?
Jon Gillham
00:20:10 – 00:20:28
Yeah, we launched with both plagiarism and A I detection at the same time. And we launched on my date a little bit wrong because it was a bit of a blur. But we launched on Friday and made that open to the world. And then Chat GP T launched on on the Monday, but it was basically launched a few days after we launched Chat CBT launched,
Jeff Bullas
00:20:29 – 00:20:32
right? So was that good timing?
Jon Gillham
00:20:34 – 00:21:12
So at first I was pissed um because we just trained the way that these A I detectors work is you build a data set on the large language model that you want to learn from. And so we were using GPT three and open A is an API playground to build a data set to train on and then um trained a detector to be able to recognize that type of content. And then what we thought initially, I thought G GP T was gonna be this totally new type of content. But then what we quickly learned was that it was a wrapper around GP T three. So I, I think it would have been incredible if
Jon Gillham
00:21:13 – 00:21:47
I would have loved to go back in time. I think we missed the opportunity. Um The size of the opportunity was not what I thought. I definitely missed the opportunity and didn’t throw everything I had at it, which I should have in hindsight and become even more of the default tool at the time. Would have loved to have sort of six months of becoming the standard for A I detection. Um Ahead of chat G BT launching, that would have been phenomenal, but this was almost as phenomenal timing. So it was great timing in the end. Um would have loved a little more, a little more time.
Jeff Bullas
00:21:48 – 00:21:49
That’s both good and bad.
Jon Gillham
00:21:50 – 00:21:56
Yeah. Yeah, exactly. I think that’s always the case. I don’t think there’s any, there’s no, there’s no perfection.
Jeff Bullas
00:21:57 – 00:22:49
No, but you can have a good idea and people struggle with it for 56 years and sometimes it’s just a matter of lasting long enough to actually start winning. And, uh, because, uh, you know, you’re saying, well, here’s an opportunity to start it. But I, I think, uh, for me, I think Apple’s, you know, headset provision, whatever they call it, um, is a good idea. Damn expensive. But I think the timing’s wrong. I think they’re still 56 years out from that. Um, a mate of mine’s been wearing them a lot and he walked into a restaurant once to catch up with me and had a walking stick and I said, what’s going on? He said, ” I’m so dizzy. I’ve been using the headset. Oh no. So I think there’s some downsides of immersing yourself in a virtual world. So anyway, that’s another story. So
Jeff Bullas
00:22:50 – 00:23:23
Now the thing I’m curious about as well because I’m just that sort of guy basically, is plagiarism a real thing today. Um Does Google and the other question is, does Google Care um and content written by A I, even though it’s almost scraped the web and put it all together. Um Yeah. So number one is plagiarism, a real big thing today that uh you initially set off to do is that still a big thing?
Jon Gillham
00:23:24 – 00:23:38
So declining. Um which, which makes a lot of sense with, with this rise of A I generative A I, why would somebody copy and paste when they can just get it rewritten? Um by, by A I plagiarism is definitely declining.
Jeff Bullas
00:23:38 – 00:24:18
Yep. Ok. So then the other one is A I writing detection does Google care because I’m quite often a lot of people and, you know, myself included, use it to do research, create an outline and then do a lot of your own writing. But you do, you might use a snippet from A I content because it’s, you’re going well, that’s just a fact. that’s all I need. Um But it’s not good at writing stories, especially personal stories and quite often good writing comes from writing good stories that make a point. So, um does Google C from your experience and expertise in this area? Does Google care about A I writing?
Jon Gillham
00:24:19 – 00:24:56
Yeah. So it’s a great question that no one can answer with certainty. Um I think what we can look at is the data. Um March 4th, they took manual, there was a manual action update that predating the, the core update bit of a psyops um effort from Google to try and scare um everyone in the seo world and say that we’re going after spam and then to do on the sort of the eve of that update that had the most sort of publicity push that they have done in a long time around an update. They may took manual action against thousands of sites
Jon Gillham
00:24:57 – 00:25:48
when we, so we looked at those thousands of sites and looked at what were some commonalities amongst them. The thousands of sites that Google had d indexed on the eve of their uh core update, the thousands of the thousands of sites that they had taken action against. The vast majority of them had the vast majority of their newest content being A I generated right now. That doesn’t necessarily mean that all A I is bad. I think there’s a sort of, I think there’s, it’s reasonable to think that not all A I content is bad is spam, but all spam in 2024 is A I content. The same way that plagiarism is declined, who’s hiring writers to spam the web, they’re using A I to do that. And so was A I content. The signal. I think what we’ve seen is that websites that
Jon Gillham
00:25:49 – 00:26:39
have a massive increase in the number of posts on their site are eventually getting hammered by Google, whether that’s due to an A I detection signal within Google or whether that’s for, for other reasons, I’m not sure. Um So I think Google has this really tricky narrative to manage where they need to be an A I forward company yet. It doesn’t take too many sort of thought experiments to think. What if A is Google search results? What if Google’s search results were filled with nothing but A I content? If Google search results are filled with nothing but A I content, why would anyone go to Google? Why wouldn’t they just go to the A I that generated the answer. Um And so I think it’s an existential threat to Google. I think
Jon Gillham
00:26:40 – 00:27:09
That’s pretty obvious. I think Google has taken clear action and their narrative needs to be A I friendly A I forward because that’s what the capital markets want them to be. Um So all that being said, I, I think there is, I think there’s uncertainty but I think most people would view publishing A I content as riskier than publishing human content if they are concerned about how Google will treat their site.
Jeff Bullas
00:27:09 – 00:27:54
Yeah. Yeah. It, it’s a very interesting game and what we have now, which um is a question I continue to ask and I’d be interested in your thoughts is that I sit in a, in my sunroom reading room uh in the mornings, reading, doing a lot of reading, taking notes quite often. I have a question I wanna ask and I’ll go and do ask Google and I have to go sift through a lot of stuff or do I just put in a well structured good prompt into Chat GP T which will give me the answer and you know, it’s, of course it’s hallucinating and you gotta check the facts. But in the main, it’s, you know, pretty, usually, pretty correct. So I’m using Chat GOT a lot now instead of Google,
Jeff Bullas
00:27:56 – 00:28:37
um I’m using to summarize, you know, a book that was written four or five years ago? Give summarize this with key takeaways in 1000 words? Brilliant. OK. So, so are you, are you seeing more and more people using chat G BT? And, and your experience too versus using Google? And is Google? And the big question is, is Google’s $100 billion plus a year business from paid ads on Google search, is that under threat? So the first question is um what are you seeing in terms of use of Google versus Chat G BT in terms of finding information?
Jon Gillham
00:28:38 – 00:28:51
Yeah, I think um I think that the impact has not been significant yet to Google which, which is, which is interesting there. The data that comes out of Google shows that their search volume is is so
Jeff Bullas
00:28:53 – 00:28:54
I think it’s Yeah.
Jon Gillham
00:28:54 – 00:29:41
Yeah. And the growth rate hasn’t. Chad GB D hasn’t seemed to have created a significant impact on um Google search volume. Yeah. Um Now it would be very interesting to try and break that down by the type of query further. I mean, I don’t, we don’t see all that data, but that would be really fascinating. Um But I think, and then, but it’s impossible to ignore the growth of CC BT and the number of people that are increasing, using it in, in, in an increasing amount. Um So I think, I think, and then they, and they have um I think they view, they view this as an as a clear threat to their business
Jon Gillham
00:29:42 – 00:30:00
and don’t, uh, and could so easily become the Kodak of the, of this era of the internet where they invented a lot of the fundamental technologies just didn’t have the, the sort of the, the business appetite to release something as flawed,
Jon Gillham
00:30:03 – 00:30:09
a perennial game of catch up. Um, and being a little bit more conservative than, than even I GP DS.
Jeff Bullas
00:30:10 – 00:30:55
Yeah. Yeah. And then we have on the, on the sidelines leaning in because they’re basically the major investor in open A I is Microsoft and Bing, which are trying to lease, you know, are basically a, a hybrid version of search search plus, you know, a chat G BT type of interface. So you got both running I, I think the challenge is this is a very long game and we might, it might take 10 years to see whether the disruption is going to happen. Um I don’t know, it would be interesting to go back and look at history and go. OK. How long did it take for Kodak to realize it was actually stuffed and about to become extinct. Well, close to extinction. It’s not extinct, but you might as well say it is. So
Jeff Bullas
00:30:57 – 00:31:34
I think that we also are also in the habit of using Google and that uh we’re also using Chat G BT. So Google, we’re still using Google. We’re also using Chat G BT as a backup as a sidekick maybe is one way to describe it. So, and, and a research assistant. So I’m using both. I actually use Google to verify facts much as Google still gets facts wrong. And, but it’s another way of verifying chat G BT, which I do regularly, especially when you get into doing a blog post on the 30 stats of, you know, youtube or something. So, um
Jeff Bullas
00:31:35 – 00:31:51
So are you seeing any trends within your industry, like plagiarism checking is dropping because people are doing something written by chat GP T and adding their own nuance to that. Are there any other trends you’re seeing within A I that you’re finding interesting?
Jon Gillham
00:31:52 – 00:32:29
Yes, we’re seeing um and related specifically to Google. So we have an ongoing study where we look every month at the amount of the search results in Google that are A I generated. And so this question sort of is A I suppressing is Google suppressing A I content. Um So what we’re trending, what we’re tracking is the amount of A I content that shows up in the top 20 search results across hundreds of different keywords. Um We’ve seen it grow from pre G BT, this sort of it was about 5%. Um Our detector has sort of a 3% 2% percent, 3% false positive rate.
Jon Gillham
00:32:29 – 00:32:53
And so some of that was probably human content being incorrectly identified as A I content. But then what we’ve seen is this significant increase since the release of Chat GP T and right now it’s sort of been holding steady over the last few months at around 14%. So around 14% of the internet, excuse me, 14% of Google search results are uh currently showing up as A I generated content and
Jeff Bullas
00:32:54 – 00:32:57
That is significant.
Jon Gillham
00:32:57 – 00:33:31
Yeah, it’s a, it’s a meaningful amount. I mean, I’d be, I’d be very curious to know, to do some analysis around like what, what percent of the internet, what percent of content that has been published in the last two years has been A I versus human. I think it would be a lot more than 14%. Um Just because of how easy it is to spam. Um But yeah, right now we’re seeing 14% of the search results are, have been, have been A I generated search results. And at what point at what point do we end up with um you know, snake eating its own tails in terms of the um
Jon Gillham
00:33:32 – 00:33:56
A I companies train on that data and then just train on their own A I data. And then we have this sort of the dead internet um theory um play out at what point does that 14% increase? Do those search results become worse? And then more people will switch to an alternative method of finding information. Um Yeah, it’s fascinating to see how that, how that’s gonna play out
Jeff Bullas
00:33:57 – 00:34:37
Yeah, you’re right. But I like the analogy of a snake eating its tail. The reality is that if the A I becomes better and better actually creates better content and then the amount of A I in the search results starts to go up to 4050 per cent. They’re going well, the human writers going, what am I here for? Which is already everyone, most serious writers are actually asking that question or not serious, but people that have produced meaningful content at scale. Um So this is the whole thing, man versus machine uh scenario, isn’t it? Because A I is actually asking big questions of what it means to be human,
Jon Gillham
00:34:38 – 00:35:25
big time. I mean, no, I certainly did not set out to build a societally important company. And I mean, I think that’s a little bit grandiose or, you know, not, not quite there. But I mean, this is the tool being used by people for some very important tasks. And we’re very squarely in this debate around both the, you know, there’s the ethical ethical development of A I and how do we not introduce bias when we’re building large language models? But there’s also sort of this really interesting societal debate around the ethical use of A I. I think there’s, you know, we’ve talked about this a lot around the lens of how AI content is viewed in the eyes of Google. Um But how do you feel about reading a review for
Jon Gillham
00:35:26 – 00:35:58
baby formula that was I generated as opposed to human generated. I think this sort of society is still wrestling with, where is it? OK? For A I content to exist and to not exist. Summarizing, summarizing a sports game and providing me the score. Great. That’s why I’m good with that. As long as the information is accurate, personal review of a product that could do harm to, to, you know, loved one’s health. I don’t, I don’t want to read an A I generated piece of content for that. I want to read. No one knows who wrote that.
Jeff Bullas
00:35:59 – 00:36:57
Yeah, I’m reading the most fascinating book in that sort of area by Mustafa. No, not sorry. Y Noah Harari, he wrote SAPIEN. I saw an interview between him and Mustafa Suleiman a while back. Um Mustafa Suleiman is also philosophical but also builds stuff in A I and uh you know, Bahan, uh deep brain, I think it was. Um So, so he’s got a book out called Nexus, which is really quite fascinating, really uh written by, it’s just been released in the last month and I’m about halfway halfway through it. But are we creating one of the big questions he poses? Uh and he actually uses a parallel to religion in that um we have, you know, we all in religion essentially, you have a God who’s all knowing infallible and also mysterious, right? That can answer your questions and your prayers, right?
Jeff Bullas
00:36:58 – 00:37:54
And if you have a look at the parallel to a IA I is mysterious and also it can be seen to be infallible. It actually answers big, big questions if you prompt, put the prompt in. Um So we’ve got this, but there’s no checks like who’s watching the machine, right? Um Essentially religion grew up in, you know, in fiefdoms and kingdoms where the king and the pope, for example, were, you know, the king and the society was a check on the religious growth of religiosity and spirituality. Whereas A I um is how do we monitor this oracle that most of us have no idea who actually gets the answer, right? Is it infallible and that’s being seen? So we have this interesting um who’s going to manage the machine and provide the guardrails.
Jeff Bullas
00:37:55 – 00:38:07
Um And we’re all wrestling with this, aren’t we? And I’m interested in your thoughts on how do we manage the A I and how make sure it’s actually working for mankind? And uh how do you put guard rails in place?
Jon Gillham
00:38:07 – 00:38:32
Yeah, I mean, these are, these are fundamental questions that definitely outstrip my, my capability um in terms of, of being able to uh uh I haven’t read that book yet, but it’s on the list. And I um if you like that book, I think uh Suleiman’s uh his own book of the coming wave. I finished it a couple months back and it’s quite good as well. Um, but it’s, I think,
Jeff Bullas
00:38:32 – 00:38:42
yeah, I read that with, yeah, I read that as well. So it’s, yeah, I just put out, I just, I just, I just put out a list of the top 10 A I books I’ve read in the last 12 months.
Jon Gillham
00:38:43 – 00:38:44
I’m gonna have a look at that.
Jeff Bullas
00:38:44 – 00:39:18
Yeah. So, uh, philosophical to practical. Um, I interviewed Nick Bostrom, uh, the polymath out of Oxford. Ah and he’s written two books, Dystopian one, which is superintelligence in 2014, the latest one, Deep Utopia. Um that, that was the most fascinating book in terms of pushing the boundaries of where I could go. But um yeah, it’s, yeah. So how do we manage A I um does the government need to step in? I think
Jon Gillham
00:39:20 – 00:40:19
so I think I, I where so in terms of like content, authenticity and the world that I that I’ve been looking at studying and understanding um I think the societal harm that can come from made up words is marginal compared to the societal harm that can come from um more acute forms of media, images, audio video, those have the ability to, to move a populace to action um in a way that I think words, um words don’t. And so what, where I think we’re gonna see um sort of regulations is around watermarking of the more harmful types of potentially harmful types of media audio video images. I don’t think text is going to get the same treatment in part because it’s the genie is very much already out of the bottle where sort of
Jon Gillham
00:40:19 – 00:41:14
open A I enforced water marking on their own models, you could just get a different large language model that didn’t have the same watermarking algorithm to rewrite it and it would totally remove the watermark. And so I think, and, and vice versa for any company. So unless, and then now with, with open source models already out, um even if that became a a requirement, um they all the, you know, Mistral, um Llama, all of these open source models would be there to be rewrites and removers of any of any watermarking. Um So I think that’s, I think the government is gonna, my guess is the government is going to step in and it’s going to be on more potentially harmful forms of content production which will be audio video images as opposed to text.
Jeff Bullas
00:41:14 – 00:42:10
Yeah, it’s a very fascinating area and I think the biggest learning um out of history and I’m calling very recent history is the use of algorithms on social media platforms such as Facebook and those algorithms are used to increase what we call engagement. In other words, the algorithm is trained. If this type of content creates engagement, we want more engagement. So what they do and we’ve discovered that, that uh those algorithms which learn from each other, which learn from a self um decides that OK, so if I create more rage and anger, I’m actually gonna get more engagement. And what’s happened is social media has been a social experiment. And as we’ve discovered, it’s actually threatening even things like democracy, it’s threatening truth, it’s threatening information. What is it? Disinformation?
Jeff Bullas
00:42:10 – 00:43:03
So the reality is that there was a really fascinating and I’ve read it a few times now is that Facebook uh went into Myanmar and launched its platform to like a Greenfield site and essentially the rage they created because the algorithms actually gave more priority and put them to the top of the page which then actually created turned the different tribes and religions against each other in Myanmar. And it was horrific and Facebook did not, it stopped it for a while and then the country became more peaceful, they’d turn it back on again and the country started turning on itself. So we have recent history showing that if algorithms and technology are allowed to run riot without what I call self correcting mechanisms, which could be the government, whatever authorities, independent, independent organizations.
Jeff Bullas
00:43:03 – 00:43:13
So for me, I think the lesson is sitting in the social media experiment that we never predicted would do this. Did we back in 2008 9?
Jon Gillham
00:43:14 – 00:43:41
No, I mean, we we I’d say, I think there was probably back then. There was some concern around the mental well being for image and like the harm that comes from like the comparison game but in terms of moving societies, um I think no one, no one predicts. I I heard no predictions of that. That’s for sure. I think that the scale at which it could move things was greatly underestimated.
Jeff Bullas
00:43:42 – 00:44:32
So I think there’s some learning from that and uh Nexus actually touches on that and I read quite a few others. And, uh, so for me, I was a social media influencer back in the day, back in the up to 2015, 16, even beyond. But I pulled back from that one. There was a personal incident that happened in my life and I didn’t want to be public for a while. But also the other part of that was, um, that I saw what going through a news feed on Facebook did to my thinking about myself. And I think I’ve got a reasonably good self esteem but you put, but then you have a young impressionable teenager or sub teen now. And the thing I really struggled with. And it came to the fore when I spoke at a round table in,
Jeff Bullas
00:44:33 – 00:45:03
uh Egypt, um which was, you know, done at Sharm El Sheikh and that was the pros and cons of social media that was 64 and six against, hosted by the President of Egypt. It was a discussion about the pros and cons of social media. And for me, that was a bit of a wake up call in that there is a dark side to technology because social media for me has given me attention, which is lovely. I, you know, we all lack a bit of attention and affirmation and actually feeds creativity and motivation.
Jeff Bullas
00:45:04 – 00:45:33
Um But the thing and there’s a phrase I remember that was shared with me a while back. It is when we go to social media and we watch people’s lives, we see the polished outside of others and we judge our insides by the outside of others. And you’re going, oh my God, my colleagues are speaking at all these conferences. They’re doing this and they’re doing that. Um I felt less and I went, I’m so I hardly ever scroll. I really don’t because I know what the algorithm’s doing to me.
Jon Gillham
00:45:36 – 00:46:30
Yeah. No, it uh I, I’m incredibly cautious. I’m incredibly scared of the impact that it has. I’ve got kids 779 and 11. Um And I’m trying to view it in a positive way. Um So that they don’t, certainly don’t have social profiles of themselves, but their youtube Shorts algorithm or the, you know, the they’re non tiktok, but the youtube algorithm gets, gets tuned. Um And, you know, there’s, there’s despite, despite my best efforts of going in there and trying to tune the algorithm. So I don’t wanna force something into their algorithm. But if they’re interested in, um, messy highlights. I would much rather them see a lot of that and some engineering questions. Sure, grades and science. But, and then turn down anything that I, I view, could be harmful to them. And even with sort of,
Jon Gillham
00:46:30 – 00:47:24
you know, some Andrew Kate stuff got into my 11 year old son’s, um, stream then none of the, like I’d say, most aggressive wording. But, um, some stuff that I think would be, would be harmful. Um And I wouldn’t want his views to reflect those. Um And so we talked about it and we’re sort of doing it with this process with them, but sort of downvoting those potentially most harmful uh items and upvoting the ones and trying to sort of tune the algorithm to provide mostly good information. I think the consumption of knowledge, consumption of information I think is, is, you know, what has separated humanity. And I’m excited to kind of read that book. Um But the sort of that, that process of um continued learning and consuming knowledge is, is so fundamental to the human experience, but it gets so um
Jon Gillham
00:47:25 – 00:48:13
so screwed up with social media and, and so anything that we can do to try and influence that. And I think A I is at the same risk of um when they’re producing a truth and it’s a single truth as opposed to the messiness that is social media and the current information world around what is true and not true. But right now, if we get to the point where we have a single brain producing a single answer around a, a statement on, of, or opinion around a type of government or, um, sort of what, what society norms should look like. Um, I think that’s, that’s sort of, you know, you end up with ideas. Not, I think that ultimately it could lead to a significant decline in, um, knowledge generation because it will reduce the amount that competing ideas compete.
Jeff Bullas
00:48:14 – 00:49:02
Yep. Yeah. And there’s so many things in this box, a genie that we’ve created and, um, with social media, it’s almost like inviting good and evil into the house and you have no choice on what’s going to show up in that house. Yeah. And that’s what’s really scary for me. And, uh, and the reality too, that really blows my mind and I still don’t get it. About 10 years ago. I was asked that, you know, have you heard that, you know, young people are getting most of the news from social media? So they’re not going to the New York Times or CNN or, you know, whatever or BBC or ABC News in Australia. Um, they’re going straight to social media and we’ve been using social media and going, uh, that’s, that’s extremely dangerous because
Jeff Bullas
00:49:03 – 00:49:47
The news that sells is the one that has the highest engagement that gets the most views. And the one that’s actually the most outrageous. So, we’re just, and we’re just viewing superficial shit most of the time. It’s actually not good content. Um, anyway, yeah, for me, the more I read about this, I’m, I’m a beneficiary of social media. The reason I built my business was because I grew a Twitter following of half a million. Now, it’s just turned into a cesspool. So there is, there is jewels in there, but it’s harder to find. Right? And for me, the joy of Twitter when it came out, was it connected tribes around the world that were interested in social media or whatever? And the algorithm of engagement hadn’t been
Jeff Bullas
00:49:48 – 00:50:11
implemented yet. So what happened was you actually found each other on tribes organically and it worked, it was brilliant and it worked. So we, it’s a, we’ve got to ask better questions. The A I um act has been denied in California in the last few hours.
Jon Gillham
00:50:11 – 00:50:16
Yeah. Saw that. Did you see that on social media or did you see that on a news feed?
Jeff Bullas
00:50:16 – 00:51:07
No, I go straight to news sites myself. I go to the New York Times. I go to CNN. That’s my thing. I go to the BBC. I go to the ABC in Australia. I try to avoid what are called sensational type news sites, but I go straight to the source. I am not going to let social media tell me what I’m gonna view. Um And I sit on my, and during, on a Sunday I actually buy real newspapers like the Financial Times Weekend from the UK and Asia. But what I love about it is its natural discovery. There’s no algorithms there. I actually stumbled across information that I will find of interest. So this is the conundrum. We have a great technology but we didn’t know that what it was gonna do because
Jeff Bullas
00:51:08 – 00:51:15
it’s allowed to just run riot with Hadley without any oversight. And now we’re trying to get that genie back in the bottle and it’s almost impossible.
Jon Gillham
00:51:17 – 00:51:45
Yeah. II, I agree. Um I know it’s gonna, and do you think, do you think the goods the, the, like the positives around? I mean, we sort of all over the place with the social media and a I, do you think the positives of social media around the connection of the tribes and the sort of niche knowledge being able to be shared? Um those positives still exist or they just overwhelmed by the, by the, the wave of sensational um nativity.
Jeff Bullas
00:51:45 – 00:52:28
Good question. I really think that the positives still very much exist. Um There’s a lot of people that go in, they’re looking for good content, but it’s on linkedin and there’s some places like linkedin, which I think have better guard rails, um more professional. In other words, there’s more grownups in the room. Um So for me, I just don’t get that. How you make, this is social media and basically bullies, you know, teenagers bully each other on social media and, and the most terrible comments are made about people, whether it’s an actress or an actor or, you know, a politician, people go in and lean in with rage and horrific comments because they’re invisible. The village would have pulled him into L A century ago.
Jeff Bullas
00:52:30 – 00:53:14
Yeah. No, I agreed. But anyway, we look, I think it’s fantastic technology. I love it and I’m curious about it and uh it’s been great. So, but let’s back to originally dot A I, we’ve gone on a little bit of an A I journey, social media algorithm, witch hunt. And that’s another story as well. Witch hunts in religion and infallibility and stories. Anyway, that’s in Nexus as well. By the way, if you wanna look, you’ll love it. Um So originally A I, what’s your next few steps without giving away your trade secrets? Where are you leaning into plagiarism? Obviously dropping away. Where are you guys leaning into where you see the opportunities in A I for your platform? Origin Originality dot A I?
Jon Gillham
00:53:15 – 00:53:57
Yeah. So I’d say even, you know, if it was ethical and legal, um I, you know, if we had a room filled with monkeys that were able to tell the difference between A I and human content, we’d have that. So I think we’re like a solution first to focus on copy editors. So people that receive text written text from somebody that they employ whether that be right, um whomever journalists and then the steps that they need to do, they need to make sure it’s not A I written, if that’s not in their policy, not plagiarize fact checking, make sure the readability score is proper grammar and spelling. You know, a lot of attention right now is around our tool and A I detection. But we’re building out a complete set of solutions that,
Jon Gillham
00:53:57 – 00:54:29
you know, whether or not your pro A I content or anti A I content. No one wants to publish factually inaccurate information. And so we have a, we’ve built a beta test version of our, of our, a fact checker and we’re, we’re about to launch the, the improved version of our, of our fact checker. Um And so basically any step that a copy editor, does anyone that acts as a copy editor, somebody that receives texts from a writer needs to do a final Q & A QC check on that content before they can hit publish with confidence and with integrity. That’s, that’s what we’re, we’re the functionality we’re building out,
Jeff Bullas
00:54:30 – 00:55:19
right. That’s awesome actually, because uh fact checking is actually very, very important. That’s one of the things that I find that with chat G BT, I’ll read something and I’ve checked it before and found it to be inaccurate. Um But Google in itself is not, is sometimes not accurate as Well, so we, we have, you know, what, what is the uh where’s truth really lie, right? And, the reality is that the truth is actually not black and white. In fact, people’s memories are imperfect. So what a person thinks is true can actually be nuanced or forgotten or substituted. So, it is really cool to do fact checking. Um Are there any other areas that you’re leaning into that are um that you’re seeing as trends in the A I content creation area?
Jon Gillham
00:55:20 – 00:55:51
Yeah. Uh uh Sort of a modern approach around content optimization is something that we have in the works. Um So sort of classic content optimizers to try and help you rank in Google um take a um call it like a dumb math approach where just like, hey, here’s the jumble of words that should be included um where we can use, use the A I research team and take a more intelligent approach and produce a far more accurate score that people should be aiming at. So that’s something that’s in the works and then editorial compliance. So everyone has
Jon Gillham
00:55:51 – 00:56:13
editorial guidelines, but how well does this piece of content measure up against those guidelines? Um So that’s something else that we have in sort of the research research stage. Um So yeah, anything that sort of ultimately, it lets somebody, it would be irresponsible, somebody to publish content on the web and not have run it through the checks on originality. That, um, that makes sense for them.
Jeff Bullas
00:56:14 – 00:56:34
Cool. So, essentially what you’re doing is you’re leaning to optimizing content in an A I world. Yeah. Cool. Yeah, exactly. Yep, Jon, it’s been fascinating having a chat mate and, um, and I don’t know whether you use the, the term mate in Canada or the word buddy. Which word do you use? Basically
Jon Gillham
00:56:35 – 00:56:46
both. Um, I think ma ma is definitely uh uh a sort of Aussie Aussie UK. Um But I think uh yeah, the definitely, definitely used.
Jeff Bullas
00:56:48 – 00:57:14
Cool. So this is one question I ask most of my guests, someone’s forgotten, but if you had all the money in the world, Jon, what would you do every day? That would bring you deep joy and I’m talking about meaningful joy and frivolous joy. In other words, what would really make you happy if you had all the money in the world, what would you do every day? It might be two or three things that come together. But what would you do every day if you had all the money in the world?
Jon Gillham
00:57:16 – 00:57:37
Yeah, times coaching, coaching, youth sports always, always uh outrageously difficult and time intensive and, and painful at times. But I think uh yeah, uh uh super, super reward. Never, never a time where I’ve regretted putting in energy into that activity and brings lots of, lots of fulfillment.
Jeff Bullas
00:57:38 – 00:57:41
Ok, great. So in other words, anything that’s not a I
Jon Gillham
00:57:42 – 00:58:02
definitely definitely. No, although I have used A I multiple times to help me build. I’ve used a CBT to help me design uh like training plans like dry, dry land trading. Ok. I got this length of time. I need to work on this build me a build me a plan and it was pretty good. Did I modify it? But definitely uh definitely a time saver.
Jeff Bullas
00:58:02 – 00:58:30
Cool. Yeah, I think that’s the right way to use it. That’s great. Thank you very much. Uh Jon, it’s been an absolute pleasure mate. And um, I’m sure if they go to your, you know, originality dot A I site, they’ll be able to contact you guys. So, um, thank you for what you’re doing. It sounds like it’s, it’s been a fabulous journey and um, yeah, the future, the future is going to be interesting to say the least. So
Jon Gillham
00:58:31 – 00:58:35
it is.