ChatGPT for SEO

40 min
Guest:
None
Episode
67
This week we're adding to the noise about ChatGPT with an episode covering some of the interesting ways we have been using it for SEO.
Connect with Michael:
on Twitter @servicescaling
on Instagram @cos71n
on Linkedin
his personal website.

Connect with Arthur:
his personal website
on LinkedIn

Watch our YouTube:
We're posting @watchtheseoshow

Our SEO agency:
Check out our agency Local Digital
Follow our agency Local Digital on Instagram @localdigitalco
Check out our content on Youtube

Show Notes

In this episode of The SEO Show, Michael and Arthur kick off the new year with an engaging discussion about the buzz surrounding ChatGPT and its implications for the SEO landscape. After a brief introduction and a warm welcome back to their listeners, the hosts dive into the current hype surrounding ChatGPT, a chatbot built on OpenAI's GPT-3 model. They share their thoughts on the excitement it has generated across various industries, including marketing and programming, while also expressing skepticism about the notion that it could replace traditional search engines like Google.

The conversation highlights the practical applications of ChatGPT in the realm of SEO. Michael shares his experience using the tool to create suburb pages for a local business, demonstrating how it can quickly generate lists of suburbs, SEO-optimised title tags, and even 250-word descriptions for each suburb. Arthur adds that while the output may be somewhat bland, it serves as a solid foundation that can be refined later.

As they explore the potential impact of AI on copywriting, the hosts discuss the possibility of watermarking AI-generated content to help Google identify and manage it. They emphasise the importance of human oversight in ensuring that AI-generated content is high-quality and engaging for readers.

The episode also touches on other practical uses of ChatGPT, such as generating meta descriptions, writing htaccess rules, and creating FAQs. Michael shares how the tool can save time and streamline processes that would typically require extensive manual effort. They also discuss the limitations of AI, cautioning listeners against over-reliance on it for complex tasks.

Towards the end of the episode, Michael and Arthur speculate on the future of AI in business, including the potential for more human-like chatbots and the implications of AI-generated content in various industries. They conclude by encouraging listeners to explore ChatGPT and its capabilities while remaining mindful of its limitations.

Join us for this insightful episode as we navigate the evolving landscape of SEO and AI, and discover how tools like ChatGPT can enhance your digital marketing efforts. Happy SEOing!

00:00:00 - Introduction to the SEO Show
00:00:39 - Happy New Year and Podcast Plans
00:01:54 - ChatGPT: The Current Buzz
00:02:45 - Understanding ChatGPT
00:04:24 - ChatGPT's Popularity and Costs
00:05:25 - ChatGPT vs. Google: The Debate
00:06:58 - The Hype Around ChatGPT
00:07:21 - Using ChatGPT for Suburb Pages
00:10:21 - Impact on Copywriting Services
00:11:24 - Watermarking and AI Content Detection
00:12:57 - Fine-Tuning AI-Generated Content
00:13:40 - Using ChatGPT for Meta Descriptions
00:16:19 - Technical Uses: Regex and htaccess
00:19:09 - ChatGPT for Debugging and Coding
00:20:22 - Generating FAQs with ChatGPT
00:22:13 - Content Idea Generation
00:23:37 - Future of ChatGPT and AI
00:24:41 - ChatGPT for Social Media Content
00:26:16 - Creative Uses of ChatGPT
00:27:43 - AI in Customer Service
00:30:12 - ChatGPT as a Tool, Not a Replacement
00:32:09 - The Future of AI and SEO
00:34:40 - Potential Misuses of ChatGPT
00:37:41 - Exploring DALL-E 2
00:38:38 - Conclusion and Wrap-Up

Transcript

MICHAEL:
Hi guys, Michael here. Do you want a second opinion on your SEO? Head to theseoshow.co and hit the link in the header. We'll take a look under the hood at your SEO, your competitors and your market and tell you how you can improve. All right, let's get into the show.

INTRO: It's time for the SEO show where a couple of nerds talk search engine optimization so you can learn to compete in Google and grow your business online. Now here's your hosts, Michael and Arthur.

MICHAEL: Hello and welcome to the first episode of the SEO show for 2023. Happy New Year. What are you doing? First episode of the year and we've cooked the intro.

ARTHUR: No, we haven't.

MICHAEL: Keep going. I'm doing okay. I'm doing okay. How you doing?

ARTHUR: Yeah, I'm good. Excited to be podcasting again. We've had a bit of a break just over a month or just under a month. Um, but keen to get cracking and have new exciting guests and new exciting topics and even have some of the SEO team run some of the podcasts in 2023.

MICHAEL: That will be happening. Absolutely. We, you may not know this, but we have about six guests lined up already.

ARTHUR: I know this. You've told me.

MICHAEL: Okay. Yeah.

ARTHUR: I don't know who they are.

MICHAEL: And then we also actually have a in-person, our first guest in-person that's not someone in our team. So very exciting times around here. Nice. I'll tell you what else is exciting. Actually, it's not exciting. It's almost annoying at the moment, the amount that's being spoken about and gone on about on Twitter, mainly on Twitter. Every second post you see is about this. What is it?

ARTHUR: Are you asking me? Yeah, I'm asking you. That's our Chat GPT.

MICHAEL: Chat GPT. I'm sick of those words, but we're going to add to the noise and do a whole episode about it today. Um, cause it is making a bit of noise in the SEO world or not even the SEO world, the, I guess, general marketing world, programming world, digital world. Um, we're going to look at it from an SEO point of view today because we've been playing around with it and have, you know, been using it to do some pretty cool stuff, save some time. Um, I don't know. People are very, very excited about this tool, probably more so than they should be. I don't know. I feel like a lot of people that haven't played with AI at all are now just playing with it for the first time and having their minds blown. But yeah, it's being talked up as like the be all end all that's going to put people out of work. I don't think it's quite that, but it is something that's pretty cool for saving time and making your life easier as an SEO. Would you agree with that?

ARTHUR: Yeah, I agree with all of that. There's still a long way to go, I think, but you know, the way I've used it, it's been super helpful for not only SEO, but even just general queries and responding to things. And it's just, yeah, it's cool.

MICHAEL: We've even used it to give different options for replies to a message from a client, just, you know, a funny one, a sassy one, or a professional one. Anyway, before we get into that, maybe we should explain what it is real quick. I'm sure most people listening will have been exposed to it already. ChatGPT. It is basically just a Chat bot that's built on top of GPT-3. We've spoken about GPT-3 on the podcast in the past. It's by OpenAI and it's, I guess, a natural language model. We actually use it in our tool, Natch, to rewrite content. You give it a prompt and it rewrites it. It or do new versions of copy. What ChatGPT is, they've basically put a Chatbot on top. So now, instead of having to, I guess, use APIs and software to talk with the AI, you can just use this interface where you can ask questions to prompt it, and it will reply to you. And the reason it's good is it's been trained by human trainers, so they've got it really good at understanding and answering questions. Anything to add to that, Arthur, from your side?

ARTHUR: No, I think you nailed it. Really, that's all it is, right? It's just a Chatbot linked to AI and it just spits out, you put in a query and it'll spit out an answer in its simplest form.

MICHAEL: And it's in beta form. Just your playing with your mic there is making a lot of noise. Sorry. That GPT would not be impressed with that. Anyway, it's in beta at the moment, so people can log in and use it free. I've read that it's costing OpenAI $3 million a day to run the tool at the moment because of the amount of use it's getting. It was released back in November, run absolutely nuts with people using it for everything. Every second post on Twitter, as I said, is people sharing some boring Chat GPT query.

ARTHUR: It's not surprising though. I mean, it's definitely, you know, I only heard about it in December and then people started talking about it. We started talking about it in group Chats and now the whole office is talking about it and playing around with it. As of yesterday, it was super slow and super laggy. It took forever to respond to anything. And then this morning it's down, so. Yeah.

MICHAEL: We were going to, um, probably dip in and out of it and use it while we did this potty, but we can't even use it at the moment because it's so popular. People are calling it a Google killer. Google's dead because of Chat GPT. I don't agree with that. I think Google have something like this, you know, many, many times more powerful and better. They're just not giving it to the public at the moment. So I don't think we should have any worry for Google dying.

ARTHUR: Yeah. Look, one's a search engine, one's a Chatbot, like two very different applications. You can't really compare them.

MICHAEL: To answer questions, you know, where people might go to Google and go, how far away is the sun? Now everyone's saying, well, Chat GPT will be used for that. Fine. I don't see how that's a Google killer because for commercial intent, like, you know, construction lawyer near me, you're not going to Chat GPT to look for that.

ARTHUR: Yeah. I think it perhaps might just be a fad, a bit of hype as well. Like, yeah, it's just, it's the thing at the moment, but in a couple of months time, you know, it will kind of fizzle out. People will just get bored of it really. I mean the people like I'll use it still to help out with work and whatnot, but I guess the general kind of hype and buzz will die down.

MICHAEL: Chat GPT, not Chat GPT, but GPT for the new models coming at some point this year as well, which will probably change the game yet again. Um, But for now, I guess there's tons of hype, as we said, like every second post I see someone going, most people don't know how to use Chat GPT, but here's seven tips on how to use it. And they're putting out eBooks and sort of promoting themselves as a Chat GPT expert, even though it's been around for like one or two months. I find amusing. We're not trying to say we're experts by any stretch of the imagination, but we have played around with it for SEO purposes. And we have a few cool things that we've done with it that we wanted to share with you guys. So maybe with all that preamble out of the way, let's jump into some of the stuff we've been doing. I'm going to start with suburb pages. I love me some suburb pages in SEO, depending on the type of business. Let's say you're a florist and you service a particular geographical area that you want your orders to be coming from. then it makes sense to create pages for those areas, those suburbs. In the past, the way we might've done that is to try and find big lists of suburbs and separate them out into regions and then create content around it. I wanted to see what it was like doing it in ChatGPT. So I logged in and the first prompt I gave it was, what suburbs are in the MacArthur region in New South Wales? And it just in front of my eyes spat out a list of how many suburbs, like 70 of them in a list. So I copy and pasted them, dropped them in a Excel doc. But then I went back to Chat GPT and said, all right, we'll write five SEO optimized title tags for an air conditioning company in Sydney. And it gave five title tags. I liked the third one. So I said, all right, take the third one and rewrite that title so that the word Sydney is replaced with all of the suburbs in the MacArthur list. And it did that. So stuff that would have involved, you know, going into Excel and doing find and replace or some sort of, you know, hackery to try and create a list of customized meta titles for suburb locations or using, um, I guess like programmatic approach in the backend of WordPress has just been done in literally 30 seconds in Chat GPT, which I thought was really cool. Then I went and said, look, write 250 words about each of these suburbs, as well as air conditioning in the area. And it went and sped out 250 words, a very dull, bland, line and length copy that talked about that suburb, landmarks in the area, things that you might see or do there, as well as air conditioning services. So stuff that you might have found out to copywriters in the past, it's doing. So now we're going to talk about whether this is good or not in a minute. But in a matter of, let's say a couple of minutes, I went from having nothing to having a list of every suburb, a customized meta title tag for it, and a blurb of copy for 70 pages that I could then in theory upload to my local business website and try and get some Google visibility.

ARTHUR: Nice. What do you reckon of that? Do you know who will be sweating at the moment? Who? Well, copywriters in general, but copywriting services, SEO copywriting services, who I have no doubt are using AI or like Jasper or even Chat GPT to write content. Yes. Um, because a lot of the stuff they write, it can be pretty shit, but you know, using this, you can get it done. Like you said, in a couple of, couple of clicks, couple of minutes, and you have, you know, SEO landing page copy for suburbs done. Yeah. So how, I don't know, like it's.

MICHAEL: So for me, the way I would use this, right. Is create all of this stuff, get it up there. And, um, get Google to crawl it, have a bit of time there so that it can settle and, you know, see if it starts ranking in the search results. And you might then go back and fine tune the pages, you know, rewrite some of the copies so that it isn't so Chat GPTE, but you don't need to do that from the start. So you sort of, you're speeding up your whole process, like crank this stuff out and then fine tune stuff once it's ranking might be one way of doing it. Yeah. Cause you were mentioning off there, there's, you noticed something about watermarking or something where people were able to detect this or even open AI potentially watermarking, right? So maybe we might as well talk about that right now. Cause that is an aspect of this. You know, we know that Google may not necessarily want everyone just using AI to trick it. So what was it that was going on with the watermarking side of things that might be a detriment to this?

ARTHUR: Well, they were discussing about basically the possibility of using cryptographic watermarking. So like digital watermarking, which is used in digital files at the moment, but it's a bit hard to do it when you're just exporting text. So if it's a file, different story, when it's text, you can just copy and paste it. So they're talking about using specific patterns or I guess, yeah, just patterns, predictable patterns in the, I guess the text that it spits out that can be detected. So.

MICHAEL: Maybe it would be like, let's say the first letter of each sentence starts with like the letters, this is written by ChatGPT or something.

ARTHUR: Yeah, something you wouldn't notice. If you're writing like long form copy, you would never even pick up on it. Google or whatever, whatever tools might pick up on it. But then there might be tools that will pick up on it that will allow you to change it to remove that watermark. So yeah, there'll be a lot of back and forth in my opinion.

MICHAEL: It's been our experience so far with AI content that you shouldn't just, like the stuff that spits out, Ultimately is not what you want to end up with like you can tweak it and change it and put your own spin on it Maybe hmm put a joke or something in the first Paragraph and that alone is usually enough to throw off a lot of these AI copy detecting tools With that thing I spoke about before suburb pages. I wouldn't even bother doing that at first This is like more about rapidly getting stuff up Indexed and then you can fine-tune it in time if it's worth your effort to

ARTHUR: Yeah, exactly. Like you should, you should never use like AI to write copy and just slap it on the website because it might work, but it's not going to read very well. Like you should always try to optimise it, make it read better for humans. But it's just a good way to kind of get ideas and the foundations and then build upon it.

MICHAEL: I'm guessing like my approach with that suburb stuff, I would have it so that, you know, all of your stuff above the fold, your, your money copy is still, you know, salesy in tune to the end user. And this is just supporting copy lower down the page that no one's ever going to read, you know, no human is anyway. It's more just for Google's benefit to get you ranking. So the fact that that is going to take a couple of minutes as opposed to hours, like it might have in the past is very cool. Coming back to that watermarking, I've been seeing a lot of people talking about how ChatGPT is just a absolute killer for schools because people are able to write a bloody essay that distinction level marks just by using ChatGPT. The real fear is that we're going to create a whole bunch of like kids who just use ChatGPT for all their tests and get results, you know, their assessment that they hand in, but they don't actually understand or know anything. So that watermarking will be really good for that because we don't want that. Like, can you imagine a world where people get all these awesome results and then you put them in front of a computer in an office and they've got no idea what they're doing?

ARTHUR: I wish this existed when I was in uni. I can tell you that. Yeah. I mean, we had plagiarism detectors. Every essay that I had to submit had to go through a plagiarism detector, but I guess this is like next level, isn't it? Yep. I was reading an article today that there's people, like there's one college student that's already created like an app that can detect Chat GPT copy. And there's other people out there doing the same. So yeah, it's only a matter of time until they find a way to crack it. But like with everything, there's going to be like ways around it. So.

MICHAEL: Well, the thing is right, Chat GPT is just on GPT-3, which has been out for a while now. Like a lot of the stuff that Chat GPT is using is, information that's even a couple of years old now, like the model's got any new, I think it's like to 2021. Yeah. It's pretty well understood. I guarantee you that Google with all of their, like the smartest engineers in the world working on this stuff, are streets ahead of where this Chat GPT is. So yeah, quite as magic as some people are making out online, but, um, No. Let's move on from an SEO point of view and have a Chat about some other stuff it does because from a saving time perspective, it really is awesome.

ARTHUR: Yeah. I've been using it for meta descriptions. Right. So basically, you can just ask it, you can plug in a couple of URLs or a URL and ask it to write the best meta description or give you options for different meta descriptions for that page. Yeah. And it'll just spit out different variations of it. And then you can, I guess, go further than that. tell it to add a call to action, make it more salesy, make it funny, you know, you can kind of really narrow it down on exactly what you're trying to achieve. And then similar to what you did, I guess, with the location pages, you can spit out hundreds of different meta descriptions for, um, different category pages on a website. So if you're working on a, let's just say working on an e-commerce site, you can just, I guess your keyword mapping can be, done in a fraction of the time that would normally take. Yeah. Which is good. You know, it's for us SEOs.

MICHAEL: Oh, I remember the days of manually writing every single meta description for a e-com website.

ARTHUR: Yeah.

MICHAEL: So do I. Hundreds and hundreds and hundreds of pages. So do I. Now you could just like, let's say Chat dbt, come up with six different meta description structures and then give it a list of keywords that correspond to all the pages and just alternate between them all so that You know, you generally have a bunch of different meta descriptions, but it's all churned out in a matter of seconds.

ARTHUR: Yeah. You could probably just upload a list of URLs, like a dozen at a time or whatever at a time and just say, write the best meta description for these URLs.

MICHAEL: Yeah. That uses a call to action to encourage a purchase.

ARTHUR: Yeah. And make it, you know, under 165 characters. Yeah. You can. Very cool. Very cool. Massive time saver.

MICHAEL: On the technical side of things as well, where I found it, is super cool is for stuff like regex, regular expression, and htaccess. So htaccess, for example, you know, when you go in there, you might use htaccess file on your server to apply redirect rules. Let's say you've migrated to a new website and you're trying to map old directories to a new directory on a website because the URLs have changed. Going into HT access and making changes is fraught with danger because if you make one mistake, it can pull your whole website down. So for an example, it might be people in WordPress. If you stuff up your HT access, you'll get a white screen. You can't log into your WordPress. You can't do anything. And for business owners that are playing around with HT access, I'm not sure how many do, but I'm sure some do, or maybe new SEOs that aren't, I guess, too sure of what they're doing. Very, very easy to break your site. I had a play around with ChatGPT and I went into it and I gave it the prompt, write an htaccess file that redirects all URLs with forward slash services forward slash in the path to the root. So what I'm saying here is I used to have pages showing on services. They don't exist anymore and they're just redirecting back to the homepage. So if anyone visits them now, they go straight to the homepage. What could have been, um, for it with danger, like typically what people will do is go search how to do that on Google, maybe copy paste stuff, put it into the HT access and break it. Um, Chat GPT has just spat out what the HT access code should be and why it's been written that way and like how it will function, which I thought was really cool. You don't have to go Google for ideas. You don't have to figure things out yourself. You just plug it into Chat GPT and Bob's your uncle.

ARTHUR: Really cool. Really cool for like coders as well. You know, debugging code, writing code. I've got a friend that does network engineering and he uses it a lot when he's like trying to solve problems with networks. So that's, yeah, really, really handy for technical things like that.

MICHAEL: Yeah. Like plug in some code and go, tell me what's wrong with this.

ARTHUR: You did, I think when we spoke about it last year, you mentioned that people were going on like Stack Overflow and writing responses to questions, but the responses were incorrect. Yes. So yeah, that can be an issue as well. People might put in too much faith into what Chat GPT tells them. Yes. And it could be a problem.

MICHAEL: Incorrect information a lot. That's sort of what's been happening. And, um, like people just, as I said, we're coming back to the hype where people just think it's just this genius, like world changing in a way it's world changing, but it's not, it's not, it's just relying on information that it's hoovered up from the internet and using these models to understand it all and give answers. So any of those inputs are wrong, then it potentially could give you wrong information. But, um, I find with more simple stuff like HT access, another one I said before is Regex. I haven't really done much with Regex for a while, back when I was sort of in the weeds doing more technical SEO audits, crawls of websites, I might have used Regex or even in Google Analytics, you might put Regex into exclusions, URLs you're trying to exclude, for example. So you're only sort of dipping your toes into using Regex here and there, generally speaking. So you're not a Regex expert, right? So generally what you'll do is go search guides or articles or, you know, how do I do this? And you might go look on Stack Overflow if they have answers there. Now, I find with Regex, you can just put in there a prompt. So you might say, give me Regex that excludes a certain path from URLs, and then you can copy and paste that into Screaming Frog so that when it's crawling, it automatically does that without you having to really think about it. And it's going to get that stuff right most of the time. It's more the, full-on coding and going really in-depth with stuff where, unless you have the knowledge to debug it, it could get you into trouble. Yeah. All right. What else? Have you used it for anything else cool in terms of technical or run side?

ARTHUR: I haven't personally used it, but looking at a couple of things that people have done on Twitter, generating FAQs. Yep. So basically just giving it the prompt to create 10 different FAQs based on a specific bit of content, and then it will just spit out those FAQs. And then you can start building out, you know, different pages, FAQ pages. You can probably build out whole content calendars based on what Chat GPT can spit out for you. So there's always that. that side of things that you can do very content based stuff, I think. Yeah.

MICHAEL: I think on the content side of things, um, one that I've played around is like a good thing for content creation is going on Quora and Reddit and seeing what people are asking. Like I know in the past for our stolen from social, you know, we just go on Quora, Reddit, Twitter, find interesting things, and then speak about it. You can sort of use Chat GPT to do that on your behalf. So you might, for example, let's say you're an accountant and you want to create some content for your blog. You could go into ChatGPT and write in, provide a list of Reddit questions about tax minimization or provide a list of Quora questions about asset protection or that sort of stuff. And it's going to use its, I guess, knowledge of what's on those platforms. I don't know how it does it, but it will spit out a whole bunch of questions. And if you're not happy with them, you can regenerate them over and over and just very rapidly build up a list of good content ideas that you could then go and create copy around without ever having to search Google or go onto those platforms and look for them.

ARTHUR: So it's a shame that we can't do it now because it's sort of an imperfect time to kind of test it out.

MICHAEL: Have you tried logging into Chat GPT while we've been on this?

ARTHUR: Yeah, I'm refreshing it now. It's still at capacity. So I don't know, it might be down for a while. It's been down all morning.

MICHAEL: Well, maybe we'll find, like surely like coming back to that three mil a day that they're spending on it, if that's to be believed, at some point, they're going to have to start charging for it or restricting use of it, right? Like it's.

ARTHUR: I thought you had to have a OpenAI account to use it for.

MICHAEL: Yeah, you do. But that's, yes, anyone can create that.

ARTHUR: So not a paid account, just, just, yeah. Cause you have to obviously log in to use it, but I thought you had to add some sort of paid account or credits or whatever to use it. Nope. That'll change.

MICHAEL: It will, like they're basically using everyone using the tool to further train the model. So I guess that's why they're doing it. And they would have a budget for it that they're happy to invest. And at some point that will stop. And then who knows, like, I guess in an ideal world, they just go to charging for it. Like they do, you know, we use GPT-3 in our tool and we just pay, you know, usage fees, basically. Like credits, right? Is it? Yeah. So it might be that, but for now it's just open slabber. You can use it as much as you want. I have used it to try and see if it's any good at spitting out tweets. So I wanted to go in there and say… I was going to suggest that to you. Yeah, it's terrible. Oh, really? Unless you want your tweets to mean absolutely like nothing. Yeah. I went in there, write a controversial tweet about… I wish I could log in and find it. I might have it open on one of my many browser tabs I've got open here. But I was just saying like, write a controversial tweet about marketing. And it came up with the most droll, generic, bland stuff that, um, I would never post. It even had little, um, hashtags in there. So yeah, it's still at capacity, but, um, yeah, I've given it multiple attempts to write a controversial one or write a humorous one. And like the humorous one might be like, why did the SEO cross the road to get to the other side of the browser or something like that? Like just not funny at all.

ARTHUR: That's pretty funny. No.

MICHAEL: I can see you posting that on your Twitter. That's not even the one it said. I made that one up. So, you know, I might be a bit better than ChatGPT as a comedian, but it's really, I personally think it's terrible for that.

ARTHUR: Yeah. I mean, it's good for like ideas, you know, like if you need ideas for anything, not just SEO in general, like what's 10 good ideas for a kid's birthday party. Yeah. Or like if you're, if you're trying to do creative, like what's a good way to kind of, you know, structure a certain thing or there's just, yeah, it's all about like idea generation.

MICHAEL: Yeah. Like headlines for my site. Like you have a headline. Yeah. Post it in and go, write me more catchy headlines in this one. So I saw an example a guy shared where he gave like five or six different titles about architecture and he just put the prompt in, make these titles more catchy. So the title he gave was the role of an architect in house additions. Very good. Chat GPT goes, here's some possible catchy alternatives. The magic of house additions, how architects transform your home. So that's much better than the one he had without any effort on his behalf.

ARTHUR: Yeah, we were like, I was using it during the, during the break to muck around with it with friends. So we were like just chilling and we were just putting in the random, like most random prompts. So like write a short story about like Scott, um, about him, like doing a diet and then whatever. And then it will spit out a ridiculous story and then you can ask it to write the story in the, I guess, perspective of a certain thing. So I did one for Robin, the SEO team about him being speed demon and it started talking about rollerblades. So then I made it, I asked her to write the story as a love story in the perspective of the roller braid. Right. And it literally just rewrote it and like, and made it a love story about a roller blade that fell in love with Rob as a speed demon. And it's just, it's just fun. Like it's just, yeah, that's awesome.

MICHAEL: That's where their three mill a day is going. People like messing around with it.

ARTHUR: Yeah.

MICHAEL: Um, I actually saw something on Twitter that I thought was very cool from this company where they had, um, they're using Chat GPT and some sort of like, um, audio AI where it can train a model of your voice. And it was synced up to deal with like big companies. So let's say Telstra have overcharged you 10 bucks. Normally you would have to call them up and deal with the pain of going through their automated prompts and speaking to someone to try and get that refunded and invest 15 to 30 minutes of your life to try and get 10 bucks back. It's not worth it. You don't do it. What this tool does is it calls, it listens to the calls and converts whatever is being said into text. So it can like understand it. And then it responds to it using your voice. So there's a little bit of a delay. If people speak, it sort of takes a couple of seconds to analyze what was said and then respond to it. But this tool was trained to chase down like overdue or like overcharged fees from banks and like, and deal with the person on the other end, keep asking, like, I want a refund. I want a refund sort of thing. And I saw this, like, it was a two minute long video where the whole call was handled by this AI on behalf of this person without them having to do it. Wow. So the person on the other end didn't even know they were talking to AI. They probably just thought they were talking to like, some slow person because the responses were slow and sort of a bit weird the way they were spoken. But he got the results. He's getting his fees back. So, wow. Who knows where that can head? We're going to end up in a world where AI is talking to other AI on behalf of us.

ARTHUR: I was literally going to ask you that. Imagine if we got ChatGPT to talk to ChatGPT, another bot, and see where that kind of ended up.

MICHAEL: Like our clients will have a ChatGPT bot talking to us with a ChatGPT bot.

ARTHUR: Maybe yeah.

MICHAEL: Humans just hiding indoors, like never, you know, it'll be purely sitting indoors, remote, no human interaction, Chat GPT doing everything while humans just watch Netflix. That's where it's headed. I hope not. I hope not. Me too. Me too. Hopefully. Where is it headed? Is Chat GPT a total game changer for the world, or is it more just a useful tool for getting some SEO stuff done a bit quicker in your opinion at the moment?

ARTHUR: ChatGPT or AI in general?

MICHAEL: ChatGPT, because that's the one that's on everyone's lips at the moment.

ARTHUR: I think it's just a tool. I think it's going to be a tool that's going to get better, but I don't think it's ever going to replace humans. It's, there's a lot of limitations still, I think, like you need to obviously input stuff. So, and the stuff that comes out isn't always perfect. So there's always going to be that kind of human element to it. You know, QC and what comes out, cross checking and making sure what comes out is correct. So yeah, it will evolve. It will get better. Um, to me it's just a super handy tool that can cut the time of tasks, you know, not even on half, but into a fraction of a time than what it would normally take.

MICHAEL: The low level, menial sort of stuff, anything like there's a lot of knowledge behind it and simple code or creating lists from data, reformatting lists of data into like, you know, columns and all that. Awesome for that. Time saver, but it's not going to replace, it's not like, Someone can go, all right, I'm replacing my marketing manager with ChatGPT.

ARTHUR: No, that will never happen. Well, never say never, but not anytime soon. Yeah. Probably not in our lifetime.

MICHAEL: It's Google's wet dream right there. No marketing managers, just plug into Google and let them do what they want with your credit card. That's where they're trying to get to. I don't know that we're anywhere near that at this point, but time will tell.

ARTHUR: It depends, you know, things evolve fast, right? Yeah. We don't know. We don't, it's hard to say for, I guess in the, in the near future, it's just going to be a tool. It will get better, but it's just going to be something that we can use to assist us and speed up things.

MICHAEL: All right. And then long-term we can use everyone's favorite line in SEO. Is it going to be a game changer? It depends.

ARTHUR: I think it is already a bit of a game changer. It is. It is.

MICHAEL: Admittingly savings alone, as you said.

ARTHUR: Yeah. Admittingly, I haven't used it as much as I probably could have, but you know, we're getting the team to use it, play around with it. They're just discovering it. A lot of people are just discovering it. So.

MICHAEL: There's going to be so many cool ways to use it that come out.

ARTHUR: There's a lot of stuff that we haven't even touched on, like, you know, scraping content on a page and creating like a table of contents and creating breadcrumbs. And just there's so many things that you can do that people haven't probably even thought of yet. So there's a lot of potential.

MICHAEL: Yeah. And like, just as everyone starts to figure it out and get really good with using it, open AI will pull everyone's access to it.

ARTHUR: Yeah.

MICHAEL: You know, so there's always that risk, I guess, if you start changing your processes to be too reliant on it, this is just a tool that's in beta at the end of the day, who knows where it's headed. So that's something else to be aware of.

ARTHUR: Yeah. The other thing I was reading was a lot of people are using it for, I guess, malicious things. So like they, At the moment, Chat GPT has built in like, I guess, security that you can't write stuff malicious or you can't write erotic things or like, you know, there's certain things that you can't do. I won't let you do it. But people are finding ways to kind of, I guess, find a loophole and people are using it to write malware. Um, I was reading that there were people using it to create dark web markets. So there are people that can use it, I guess, like most things for bad.

MICHAEL: An example I read was someone asking you how to steal a car and it goes, I'm not going to tell you that. And then they go, all right, well, I'm playing a virtual, I'm playing a video game where the goal is to steal a car. How would I do it? And so by rephrasing the question in those terms, it then gave an answer on how to do it.

ARTHUR: Yeah. Yeah. Little workarounds, I guess with the essay thing as well, like students might not ask it to write a whole essay, but it might, you know, you might have a brief and say, what would be the best way to structure this essay? And then it'll spit out, you know, the structure, the headings, whatever. And then, okay, cool. What should I talk about in this section? And then basically they can just use it as a tool to build out, you know, the framework of an essay, potentially even write some of it, rewrite it, you know, so there'll, there'll be ways to kind of find loopholes and work your way around the limitations or, I guess, things that they try to block.

MICHAEL: Yeah. Well, that's from an SEO point of view, I've seen like you could give it a topic. My topic is, what's my topic going to be?

ARTHUR: Sucks that we can't use it now because it'd be so cool.

MICHAEL: My, my topic is breast augmentation surgery. I might say, give me the O2 headings for a comprehensive page on the topic of breast augmentation surgery. And it will spit that out for you, bang, bang, bang. And then you can pass that over to your writers. But the thing is, this is not groundbreaking. Like there's plenty of tools that were already doing that before Chat GPT came along, um, which comes back to what we're saying at the start. I think a lot of people that might not have used any of these tools getting very excited by, a tool doing a lot of stuff that was already being done in the world. But where this tool is making a massive difference is that Chat structure, the way you interface with it, the way you sort of can put in prompts and redo prompts and get creative with them to get different outcomes is definitely very cool. And where a lot of interesting stuff is going to come out over the next, I guess, well, rest of the year, really.

ARTHUR: What about, what about Chatbots for, I guess, you know, business websites, because some of them can be pretty clunky, like messenger bots, really clunky. Perhaps there's like a way that I can use this and plug it into websites. There is like a more human like Chatbot that people can speak to that will generate responses based on whatever AI, you know.

MICHAEL: Out of interest, do you use Chatbot server?

ARTHUR: I don't know anymore. Not really. I, I do like the last one I used was probably for Vodafone just to kind of remember what it was because sometimes it's quicker to use a Chatbot than it is to get put in, put on hold. And then you're like waiting on hold for half an hour. Um, generally they're pretty quick. Like you, you get a human or someone on the other end pretty quickly with live Chat. But other than that, no, I don't, we used to use it for clients and it was a lot of time wasters usually.

MICHAEL: We used to use it for ourselves even. And then we stopped because just no one was using it anymore.

ARTHUR: Well, no one uses it. And the people that do, it's just time wasters. Like, you know, you have to have someone there on the other end, like responding like that. Otherwise it's going to hurt your conversion rate. It's going to piss people off. So. Yeah.

MICHAEL: And the question is like, how much, how much? Yeah.

ARTHUR: Stuff that can be, you know, Answered with an FAQ. Yeah. But I guess this is where ChatGPT can come in. And if you have like a more advanced Chatbot, it will probably be able to like take what all the questions that people are putting in there, start formulating, you know, responses. Yeah. More like human like responses. So it could be that people will be talking to Chatbots, not realizing that it is actually AI thinking it's a human on the other end.

MICHAEL: Absolutely. Well, that's already like, you know, there's Chatbots that you can train on your knowledge base. So like big companies, like your Vodafone's of the world would have very detailed knowledge bases and documentation, standard operating procedures. These AIs can just read all of that and understand it all so that when questions are coming in, they can answer it really quickly. So this is what we're saying. We're going to head to this Nirvana where there's very little human interaction, seemingly.

ARTHUR: I kind of, I want to play around with Dali, Dali 2. Like I haven't played with that at all. That looks fun.

MICHAEL: We'll play with it. And next week we'll do an episode on Delhi 2 for SEO.

ARTHUR: We can. Do we have access to it? Do we? Without open AI?

MICHAEL: I haven't really, like these ones, Chat GPT where there's immediate use for it and GPT-3 in our tool and the like, whereas the image one is more just fun. It is fun. Roller skating image of Rob, for example, and have some fun, but in terms of actual outcomes for our business, it hasn't been as immediately obvious to me the use. I guess generating like featured images for blogs and that sort of stuff at scale.

ARTHUR: I'm thinking what if you could probably use it for like creative, so not SEO, but maybe like Facebook, you know, create a banner for Facebook with this call to action, this text and this background and this person doing this. Yeah. Maybe you could use it for that.

MICHAEL: That's what we were touching on the other day, like ads will be written by AI with creative done by AI, served up by AI, selling products to hopefully people, but largely bots on display network. And it's just a whole bunch of computers talking to each other.

ARTHUR: In the metaverse.

MICHAEL: Oh, don't even get me started. All right, well, let's wrap things up. I think we've somewhat covered the topic of Chat GPT today. Go on Twitter, search that word. You will be inundated with posts. You'll probably find some of the cool stuff that we've covered here and maybe some other stuff that you find interesting. But that is all for this episode of the SEO Show. So until next time, happy SEOing. See ya.

INTRO: Thanks for listening to the SEO Show. If you like what you heard, don't forget to subscribe and leave a review wherever you get your podcasts. It will really help the show. We'll see you in the next episode.

Most recent episodes

View all Episodes