Ep 379: The rise and fall of AI content madness, with Ivana Flynn
:
Hello everyone and welcome to yet another episode of SBC iGaming Daily podcast, the Marketing Edition. iGaming Daily is presented by OptiMove, the number one CRM marketing solution for the iGaming market. And we have yet another special offer for you. OptiMove is offering new clients a free first month when you buy OptiMove. For further information and claim the free month, please go to OptiMove.com slash SBC. I am joined by two beautiful ladies, and we have a very interesting topic for you. I'm Ivana Flynn. Welcome. Hi, my name is Emma Byrne. I'm the director of publishing at Gen2, formerly known as Gig Media. Hi, my name is Samira van Diepen, and I am the founder of Content Lab. And probably from the titles you're already getting it, yes, we are going to be talking about content. Is it important? How are we dealing with it? How did it change over the last years? Which tools are out there to help you or haunt you? And that's why I have our two experts. Ladies, thank you very much for joining us. My first question is, last year went crazy with the AI content and we've seen so much going on. Did you see some bad impact on your business or your performance or how your team started to work or change their work based on the race and fall of AI tools? I have not directly see it within my business, but I do know a lot of clients raised concerns about AI and the use of AI. And within our company, we have become very aware of writers using AI and yeah, we've tried a lot to figure out tools to avoid from our side to deliver AI content. gaming companies and probably also a lot of other companies have been trying to use AI for their content. But yeah, we try to stay away from it. I think I can answer from not an agency perspective, but from like a media point of view, I think we had a kind of three different reactions. We first had panic, like firstly, is AI coming in? Is it going to take people's jobs? Is there still room for content? Secondly, we had teams who were very interested. Okay, how can AI improve my content? How can AI help me with my job? And then thirdly, you know, kind of what does this mean for the industry as a whole? And what does this mean for content in general? So it was very much, I guess, fear, understanding, and then I guess a lot of curiosity kind of around the whole AI, which I think isn't just content. I think it's AI in general. There's a lot of curiosity. But last year we were discussing this in every conference. There were so many podcasts about, is AI coming for our jobs? And I remember Samira was with me on a conference and we were discussing it. And as you have a content agency, you had a real fear that people are going to stop purchasing quality content. Did that change with all of that March update cleanup? And we know, we discussed it already on this podcast, that it cleaned up lots of AI content, not only, but a lot. So did you see that? Did you see impact on your work? can absolutely see that the fear that I think a lot of people had last year, also me obviously as a content agency, I've always believed that quality stands out and it's also something that with Content Lab we've always valued the most. So I always believed that AI would not take over, but I had a little bit of a fear. But after the big March Google update, that fear for me has left. I can see that everyone is focusing on quality even more and everyone that has tried to do something with AI has gotten back from it. And yeah, I can see that quality is something that AI is not yet able to give. And talking about AI content, I mean, obviously, it's very easy to use it, GBT or anything else you have there for English speaking markets. Is there any tool that successfully created content for other markets, such as Polish or very difficult ones, such as, I don't know, Japanese, Vietnamese? Is there a tool that can actually fake good AI content? To be honest, I think there are many tools out there that can produce quite native content in basically any language. However, in my opinion, the content is always very generic. So I wouldn't mark it as good content, but it's, yeah. So in my opinion, no, there isn't. But I think if you would ask this at 10 different people, you might get also very different answers. We've also tested a lot because we have markets from America to Australia and I think the same echoing Samira that we just haven't had or we haven't seen much success with it. I think especially around EEAT, which is a big focus in content right now, AI just isn't able to deliver what you need. But then for other niches, say like crypto or. Short news or email marketing, I think there's a lot of use or prolific use of AI and I think some of what comes out is quite good or acceptable and now obviously we see it everywhere. There are some websites that are entirely AI generated content and they're more than happy to advertise that they are AI written and they are produced by AI. Interesting. I wonder if you can actually build a successful affiliate website using only AI advertising its AI and outrank the good content that has got it or follows all the helpful content update. What do you think, Emma, you're in that in that niche? Yeah, we haven't seen it yet. But I mean, nothing is out of the realm of possibility, right? I think, you know, the SERPs are getting harder and harder to navigate. I think we're seeing a lot of different tactics being employed by competitors, operators, other affiliates. So is AI going to be the future of that? What part does AI have to play in content generation for affiliates? I think it's still to be seen or to be confirmed. But what I can say is, you know, I think to echo the song from, I think it was the 80s or the 70s, video didn't kill the radio star. And I don't think AI is going to kill content writers, content producers or content agencies. I absolutely agree. I think if you have invested in a good content and good quality of everything, not just content, you survive easier all the horrible Google updates. And lately it feels like volatility is the new normal. We have updates all the time. And if they don't tell you it's an update, just there is some volatility, the weekend volatility and Father's Day volatility. But we all try to... the AI content and it's Samira says she has got experience with lots of writers supplying AI content. How do you make sure that is not an AI content? Because there are all of these scanning tools and they fail, right? I mean, every tool marks it differently. So how do you how do you recognize that is an AI content or not? I think firstly, where it starts with us with the onboarding procedure when new writers start. They are made very aware of the fact that we do not accept AI content. But obviously we do scan and we do check. Our editors are very well trained in spotting AI content. I think AI uses a lot of, you know, competitive words throughout. throughout content and yeah, we do use a lot of tools. I have not found yet one tool that is scanning absolutely flawlessly, that doesn't exist. So we use different tools. We do aim to always have a clean AI scan, but yeah, that's not always easy. But even if you get that, I use different tool and it flags up as 80%. I mean, How do you fight against these different tools and different pronouns, the understanding what is AI? I believe that AI scan or a content, even if you write it yourself, if it's very generic, it will scan as AI. So I believe that even if a writer wrote it himself or herself and it scans very high, it probably means it's too generic and the quality is not where it should be. and it needs to have more of a personal touch. Just to point out, I think it's very interesting that we're still talking about having a human editor at the end of it. So whether you have AI content generated purposefully or not, I think even those that are actively using AI content are still having a human editor at the very end of that process, quality checking and assuring. Because like Samira says, sometimes you do get a lot of very generic content, or equally, I've seen content where it's like a thesaurus. wrote it, as in they're taking every complicated English word and just throwing it into a sentence, and it just doesn't read very naturally. So I think it's interesting that as far as AI content has come, that there is still, purposely or not, there is still a human at the very end of it either checking because they think it is AI content, or equally they're checking it because it is AI content. But is AI scanners, are they becoming a problem for you, Samira? Is it something like, client would return a content saying, sorry, it's like 3% AI. I don't believe it's written by human. Yeah. It is something that I see a lot actually into the extremist that we tend to attach a clean AI scan in each and every content piece that we deliver. So no one has, you know, no one has to get back about it. I do tell my clients always that we scan. I also tell people that you will never get a clean AI scan. So we have benchmarks of like a certain percentage of what can flag. And then it depends a bit on the client. Like I know that not everyone believes in the AI scans. So if I know they don't really believe in the scans, obviously we're a bit more flexible with it, but it still means that if we get a piece. and it scans on the different tools all above 80%, we will not submit it, we will get it rewritten. Emma, do you check it? You get content from your supplier, do you check it in some AI scanners? Yes, we also do rely on the agencies like Samira to provide us with scans from their end, but we do run our own. And then lastly, we have editors who also check from our end. So again, they're still a human at the very end of the process, but we do look at the AI scans, we do run. we're not looking for like a perfect score, but we're looking for a believable score. Exactly, we do that too. A believable score. We all know that it's impossible to get like a 100% human written article, even if it's 100% human written. But I think it's like a combination of the scans with some common sense as well, and reading the article to see if... you know, if it's more, if it's personal enough, if the quality is good enough. Cool. I mean, we're getting to the same point. Even if AI is there to help, we still are humans and we want to check it by humans because we want to provide by humans. So I hope that this trend keeps up because as you know, we are now recording in August. So, you know, keep this in mind when you're listening to this. Google is announcing yet another update. Is it something that you worried about? I mean, I always... check for these updates and I'm worried is it linked to the content, but lately, everything is on content. Every month there is at least tiny little bit weak on content and the big ones are tend to be on content. And it's like taking Bazooka to the last remaining, what Google understands is AI content. Is it something that either of you, because you have very different perspective, are worried about, are looking into preparing for that? Yeah, absolutely. I think there's always a fear element, obviously, you know, you work very hard on websites and you get them to a very nice place where they're consistently performing and then suddenly Google decides, ah, I'm going to have some fun with you guys. And then, of course, you have the sites that were hit previously, and you've done a lot of work with your SEOs, content, design, product, and tech teams to get them to a place where you're hoping that there is a Google update, because then you're hoping that, you know, maybe to be rewarded for some of the mistakes that you've made or some of the things that have been flagged in the past. So like poor core web vital scores, and you failed on the helpful content update. Sabina, how do you feel as a content provider? If this is on content, is that going to be a game changer for you or you just believe that quality wins every time and people will just know it and come back for more? I believe quality wins always. I do prepare for it because when Google announces big updates, I do see my clients panicking and they have older pages that they feel like, okay, this update is coming. We need to update these pages and this is outdated. I feel from everyone the stress and then I also see, you know, the urgency in getting things done before the update rolls out. So yeah, I do prepare for it. But yeah, as a supplier, I think it's a different way of preparing for it than for example, Emma. prepares for it. I think Emma touched on an extremely interesting point when it comes to Google rewarding websites that work on themselves to become better through the next update. Have you seen that lately? Because from helpful content update, people just dropped and dropped and dropped, and I barely see any success story of recovering that. So have you seen anything that has recovered after being hit by helpful content update? Doesn't have to be just in gig, obviously, but altogether we are. all SEOs and content nerds, and we look through the data and read other people's articles. Have you seen a success story? And if yes, what was applied to make the content good again or to make Google fall in love with your content again? For my side of what I've seen, I've seen a few success stories within the EN market, where you're kind of getting these smaller, older sites resurfacing and resurging. I think a lot of it is being reworded onto a completely new. content strategy where they are focusing on helpful content, adding all of these user generated metrics, is this content helpful? Would you like to leave a comment here? Did you find this good? Did you find this bad? Did you find what you wanted on this page? And I think, again, narrowing it down to, OK, this page is talking about subject X. Is X being addressed thoroughly from? the header all the way down to the very bottom. So is the intent or is the query being satisfied? So kind of back to basics, I think, really, in terms of content, like back to basics, good, clean content. I think that was always, and if you didn't stray away from that good content written by humans for humans, when Google said it's for humans written by whoever, I think you're fine. March update was really horrible to lots of websites. And OK, not just AI content, but AI content mainly. And it was really when you think of it lacking any eat. It was lacking any human touch. It was all conclusions and delve into and all of those triggering words. So I think in a way, it's good. Google is looking more into quality. On the other hand, we can still fake the quality. You can still fake it. You can still buy an expired LinkedIn profile and link it to you saying, you know, I work here and I write here. Or you can still create a persona and put picture there saying, this is Emma and I work for Geek and I wrote this piece, but it might not be true. So is there a way, in your opinion, to detect the real eat metrics from the not-so-real eat metrics? And if you're going to do not-so-real eat metrics, how do you do them to pass? That's a really good question. There's a lot of juicy things in that. I would say in terms of assuring some sort of realness to things, I think you also need to obviously take a look internally. Were you employing any of these phony or shadier tactics and then addressing, okay, how can we? So I think something is showing your authority on this. So that is building a real authorship with active social profiles, active mentions, not just on your own website, but also a wholesome or holistic approach to it. Hi, my name is Samira. Hi, my name is Emma. I write about this. You know, you'll also find me contributing on this website, this website, this website. These are more of my pieces, et cetera. This is what I talk about. So you won't find me publishing on something that I'm not attributed to. So if it's iGaming, then I'll talk about iGaming. If it's gardening, I'll focus on gardening. If it's something else, then it's something else. And then regarding towards, I guess, the... the way to fake it until you make it. Of course, there are a lot of things that you can do. So EAT, so of course you can add a bogus website, you can add bogus addresses, but you know, not real authors. You know, you can, I've seen some very good fake LinkedIn profiles, I think we all have, and I've also seen some very bad ones, and I've seen some that I can honestly say, I'm not too sure until I dig in, are they real or are they not? Exactly. You can buy followers, you can buy connections, you can buy everything for a few cents. So I think for a human to recognize that this is a real person or it isn't is difficult, and I don't think Google really understands that. They just check like, okay, it has email address, there is a picture, there is few followers, therefore a real human being. I mean, this is how we fake the bot traffic as well, right? We create a persona of the bot. So... Can we still, like let's say I run an affiliate website and I really don't want to put my name to it and my face to it. Can I still fake it? Can I fake enough of persona to pass it in Google? I think I can, but you're a little bit more into content and publishing. So what would you say girls? Can we, and how would you advise us if we want to go a little bit that naughty way? Yeah, what I would say obviously is I wouldn't advise it. I would just say, I mean, I understand there are certain reasons or certain markets that people don't want to have their names attached to. I think it's still important to try and find someone even so you don't have a website that's published by one person. You need different personas to talk about different things because one person is not going to be an expert on several different topics. Even with iGaming, you have various niches, crypto, sports betting, casino, games, lots, very different topics. So I think. It's fine to have, it's fine I think to maybe have one fake author that's still built up the same way, but you need to treat them as if they are real authors. So I think also bolster them with your real authors and complement the strategy. So, you know, do they have an email address? Do they have an active social media? Do they have a believable headshot? Not using, you know, this person does not exist dot com or. other photo generating software, where are they contributing? Are they mentioned in, are using them as part of your backlink strategy to build up their mentions? Where else are they published? What are their mentions like? What is their tone of voice? Is it consistent across all of their pages? So I would almost say that the amount of work you're going in to make someone look real is... the same as that you're going to spend on a real person. So then you have to ask, is it worth it or is it not worth it to you at the end of the day? Because it is a lot of work to build up an authorship strategy that works on your website, especially if they are not real. I love it. This is like building a whole new persona because you're a criminal and you're running from your past. And I think this is where we leave it. I mean... This is fantastic information, how to create fake users if you need to, obviously we put lots of thought into it. I wouldn't come up with half of that. So if anyone is looking for fake users, speak to Emma for more. And thank you ladies, it was an absolute pleasure. And I think we shared lots of fantastic gems. Thank you.