ChatGPT has taken the content world by storm. OpenAI announced the release of this, and there have been people posting videos about the new capabilities seen with this user-friendly form of text generation. Many people have been making articles and statements about the future of content writing, the future of SEO, and how this impacts the internet.
Table of Contents
Transcript
In the latest episode of Hack My Growth, we’re taking a look at AI content generation, ChatGPT, and the future of SEO. But we’re going to talk about in this video how this is not new technology, and what this means for SEO. Does it mean that we can just outsource all of our content and change everything that we’ve been doing up to this point?
Who is OpenAI?
Let’s start by asking the question, who is OpenAI? OpenAI is a nonprofit artificial intelligence research laboratory, and it’s got some pretty famous founders here. You might notice a few of the names, but their mission is to develop and promote friendly AI in a way that benefits humanity as a whole. OpenAI is focused on developing artificial general intelligence like deep learning and things of that nature, reinforced learning as well as unsupervised learning.
Now, we’re not going to get into the details of artificial intelligence in this video, but we’re talking about how it applies to us as far as SEO and marketing go. It’s good to know who’s behind the tools that we’re using so we can understand a little bit of their ethos, why they’re creating them, and what the thought process was in creating some of these different tools.
What is ChatGPT?
ChatGPT, it’s a natural language processing (NLP) chatbot and it’s built off of GPT-3 language model. GPT-3 has been out for quite some time. It’s a language model that OpenAI has built upon over different iterations of the GPT project. Now, the thing with ChatGPT is that it can engage in more of a natural conversation with users and generate these coherent responses to most queries. I mean, it’s still a computer.
One of the things we have to remember is that when we’re dealing with these types of things it’s “garbage in, garbage out.” Now it is powered by GPT-3 technology. That technology has been trained on tons of information from the web. ChatGPT can be used for a variety of different applications, from it’s customer service to conversational bots and AI. If you begin that conversation with people, you get some pretty interesting feedback and connections.
What is GPT-3?
But let’s go now to the next step. We’ve talked about ChatGPT, but what is GPT-3? This is called the Generative Pre-trained Transformer 3, and it’s an advanced natural language processing system created by OpenAI. This is the engine underneath ChatGPT. This is the system of unsupervised learning that is powering things such as human-like text, and making it a good tool for summarization, translation, or question and answer. It’s widely considered the most powerful language model right now, and it’s been used for quite some time. This isn’t something that just got released.
What people are seeing right now is the ChatGPT playground, which is something that can be used. Right now, it’s free because it’s in a research stage where they’re still training it and testing on a lot of other things. But the artificial intelligence behind it has been around for quite some time and it’s being used currently in a number of other AI applications.
Differences Between ChatGPT and GPT-3
What are some of the main differences between ChatGPT and GPT-3? Well, the main difference is that GPT-3 is a much larger, more powerful version of what you would see with ChatGPT. You’re going to get a much larger subset of trained data when you’re accessing GPT-3. It’s based on transformer architecture‚Äîif you know anything about that. Google made massive advancements when it comes to transformer architecture, like BERT or T5.
This whole idea of transformers is what has opened the world to see some of these advancements within the realm of natural language processing (NLP). GPT-3 is a fully unsupervised model. They’ve just taught it a bunch of data while ChatGPT is semi-supervised because it has a more specific purpose primarily in the form of chats and conversational applications as opposed to GPT-3, which has a wide variety of applications that can be used for.
Other Tools That Leverage GPT-3
As I said before, there are a number of tools on the market right now that are using GPT-3. Drift uses it to power automated conversations with website users. MailChimp uses it for AI content-driven recommendations. Autopilot uses it to automate customer journeys and personalized emails. Contentful uses it, again, for content recommendations. AdEspresso uses it to help power its AI-driven targeted ad engine.
As you can see, this technology is being used. It’s being used in the marketplace right now. It’s probably being used in tools on this list that you’re currently using, but you just didn’t realize it until they had this big movement. It kind of blew up in the world of social media, and all these new articles are being pumped about AI-driven content, especially as it relates to ChatGPT.
Let’s take a step back for a second. We’ve said a lot of acronyms. We’ve talked a little bit about artificial intelligence. Let’s talk about why and when we should be using this type of technology when it comes to our marketing.
Relying 100% on AI-Generated Content is a Bad Idea
Generated content can be inaccurate, misleading, and contain poor grammar and syntax. For instance, it will repeat misspellings when you put misspellings in there. Sometimes if it doesn’t have the right information, it can give you the wrong information. Like that whole idea we’ve talked about previously is garbage in, garbage out. You need to make sure that you know, and you don’t set it and forget it like a lot of people think you can do.
Another thing that you need to be aware of is that it’s generated using algorithms. It can sometimes create content that’s not relevant or useful to your reader. You can’t just push a button and post it. You have to make sure it makes sense.
AI content also can’t take into account things like nuances or cultural considerations. You can create stuff that is offensive. We have all heard of how different bots have been manipulated to do things that may have not been so good for society, or maybe just made those bots to be very culturally inappropriate in a number of ways. Again, whatever data the bot gets is what it’s going to spit out.
Lastly, it sometimes can lack creativity and emotion, and personality. That makes it difficult to engage your readers. What I’m seeing right now are all these people, (and there are a lot of people out there) saying, “This is the future of content, and will never have to hire a writer again. Look what we can do. We can just set it and forget it.”
These are just some of the things you need to take into account. You’re going to have content that’s not helpful to your readers. It could be misleading. It could cripple your authority and trust with your audience. That is not the way AI content should be used. It’s not just to generate content on its own and then you walk away smiling because you pushed a button and got something done for you.
We can’t do it this way, but with that being said, AI is meant to help us. How can we use AI for content creation? Well, it can assist in a number of ways. It can help you to research topics, generate ideas, and start your content out. If you’ve ever had brain fog‚Äîyou’ve been trying to create articles, and you’re just so tired because you’ve just been writing and writing, and you don’t know how to start‚Äîthis is a great thing to use to help you get started, to give you ideas, and to generate topics.
It can also help you suggest relevant keywords and insights into content performance. AI can be trained to do that. GPT-3 can do that. AI can also be used to identify the format of content that you should be using, blogs, videos, or infographics. It can help you cut down research time to make better decisions as well to filter through all that noise because there’s so much noise out there.
It can also be used to help optimize content. Think of tools like Frase and Jarvis that use GPT-3. We can use AI in a lot of ways to help us create content, cut through the noise, and do better research, but we can’t just push a button and let AI do everything for us.
How to Correctly Use AI in Content Creation
Do keyword research. You need to understand the intent of the terms you’re working with. You need to understand what type of content you need to be creating. You can’t just rely on AI. Use your SEO tools, use your keyword research tools, and then go to the search results themselves, and see what’s happening.
AI is very helpful for creating briefs. Once you understand the intent and the content, you can use AI to help you build a brief. Then you can use it to help you create drafts. Now, once again, don’t ask it to write the entire broad article for you. Don’t say, “Write me a thousand-word broad blog post on X.” You’re going to get junk. You need to chunk it, and you need to make sure you’re asking very specific questions if you want good answers. You can’t be broad. Remember‚Äî”garbage in, garbage out.”
Then you have to check and verify the information and stuff that you know to be true but also leverage some other tools. You want to make sure that it’s not seen as AI-generated content, but it’s content that follows along with Google guidelines. Google knows when your content is 100% AI-generated and they will not rank it. In fact, they say that it’s against their guidelines and they’ve been saying this for almost a year now.
John Mueller came out in April of 2022 saying, “Hey, we know when you’re doing this and we are going to punish that content. That content’s not going to rank.” Remember, this is built on transformers. The breakthrough in transformers came from the AI team at Google with the implementation of BERT and T5. Google is very well-versed in the world of AI. If you think they can’t understand that your content’s been created by AI, you’re crazy because they absolutely can.
The next step after you create your draft is to do human editing and auditing. You need to make sure that you add your voice, verify the information is correct, and cite references. If you’re getting information from people and you’re taking information from other websites, you have to give that credit back. You have to cite references. It’s essential to good content best practices. You cannot plagiarize other people’s content. And guess what? That can happen when using AI.
Now, GPT-3 is very powerful and you can ask it some cool things. You can even say, “Hey, tell me about this specific topic and give me some references.” It can do research for you where you can say, “I’m going to go read those four articles, and here’s a summary of those articles.” Make sure that you cite those articles if you’re using that information in your content.
Then lastly, you need to follow SEO best practices before you publish it. Make sure you have the right title tags and meta descriptions. Make sure you’re using the right content structure, internal links, and adding in schema markup where it’s applicable. If you follow these best practices, you’re going to do well.
Let me show you a quick example as we wrap up this video. This is Frase. Frase is a great tool for writing content briefs. This is not a commercial for Frase. If you’ve got another AI writing tool, use that. We tend to use Frase here at our agency. We use it for structuring content, trying to figure out what we’re writing about, what type of content is ranking, and what are some of the parameters we need to have within our content.
As you can see here, I’ve got a content brief. Now, some of this was written using the help of AI. A lot of it was the AI that Frase has built-in, which helps us. But also some of it was just based on my own experience with leveraging artificial intelligence in the world of SEO for the past three years.
Now, as we talked about, we want to use AI to help us build this structure and this formatting, and Frase can do that for us, which is awesome. Then we can take that content and start to create it. You notice this is the content that’s on the slides that we have just walked through. Almost every one of these paragraphs were generated with the assistance of GPT-3. I prefer to use GPT-3 because it’s more powerful, and I can do more things with it.
If you’ve never checked out OpenAI and opened an account and played around with it, you’d be surprised by all that you can do. It’s not just creating blog posts, and summarizing articles. It’s helping you write code. It’s helping us do a lot of things where we need assistance. Things that are repeatable‚Äîthat’s what AI is for.
Now, once we’ve done this article and spit it out, you want to check it, you want to make sure that it is accurate, and you have to do some research on your own. The tool that will help you see how accurate this is going to be is originality.ai. If I scan this article, there’s a 50/50 chance that this would be seen as an AI-detected piece, and that this piece was absolutely generated using artificial intelligence.
Now for me, I use this as an outline, as research, and then I put it in a video form. I’m already adding my own context to it. I’m adding my own spin to it. I’m adding a lot more to this video so that when I am done with this, it will not be an AI-generated piece although it was an AI-assisted piece.
Now I can take this and build a transcript. I can put it on my blog, optimize it for search, and send it to be indexed. I’ve got a good piece of content that I was able to do a ton of research on, in an area that I already have some familiarity with. I’ve been using OpenAI for a number of years, and now I can create a piece of content much faster with the assistance of AI, and I’m not allowing AI to do everything for me.
I hope you found this video helpful. AI can be very beneficial. ChatGPT is very cool. GPT-3, in my opinion, is even cooler. But there are many more things that you can do to go above and beyond. I love the world of natural language processing. I love the world of artificial intelligence, and what we’re able to do to help make our jobs flow much more seamlessly, but we can’t rely on AI to do everything for us.
One of the things that we do offer here at the agency is SEO coaching. If you’re an agency, or you’re a site owner that’s wanting to start leveraging AI, and you’re looking for best practices and coaching on how to do that specifically, I’d love to talk with you. Contact us and get a free discovery session.
I’d love to see what you’re trying to do, how you’re trying to approach it, and give you guys some best practices. You can leverage artificial intelligence to make your processes better, and do it in a way that’s going to benefit you, instead of generating a lot of pieces, and then none of it getting indexed or doing anything for you.
Again, I hope you found it helpful. If you have questions, comment below, we’d love to continue that conversation with you. Until next time, bye.
