Prepare to compete with AI-generated content
Chima says: “My advice is almost more of a warning: we are going to drown in content in 2023, because of AI.
Those of us who look at SERPs to get results, we’re going to drown in content, because there’s going to be too much of it. There is already a lot of content out there right now, but there is going to be even more as more people turn toward AI.
In 2022, Google released the helpful content update. A lot of SEOs assumed, and rightly so, that those who are depending on AI to create content are going to be the worst hit because they are relying on AI-generation tools to write the content, come up with the brief, etc. – everything short of hitting the publish button. In the short term, some of these websites will suffer but, in the long term, the tools will rise. The tools are going to get better at creating briefs, writing content, and optimising that content with semantic entities. They are going to get better at creating content that ends up on SERP.
Is it usable? Is it useful? Who knows, but we are going to see it on the SERP. As they get better at giving Google what it wants, we’re going to see AI-generated content that is scalable. It’s scalable because you can easily just type in a keyword and it generates the brief, writes the content, and you just hit publish. You can do that 20 times a day just covering one keyword and building clusters. It makes it easy to scale content marketing and attribute authority faster. This is what we’re going to keep seeing on SERPs, because they are producing content faster and at scale in a way that no one else can even dream of competing against.
Now, what does that mean for those of us in SEO and those of us who do SEO content? When we’re trying to fight AI, we can use parts of AI and figure out a way to get better than those who are relying on it completely. We can use AI to improve content optimisation; to find the most common questions or subheadings that pop up for a specific keyword, and then we can optimise our content for those questions and those subheadings. That way, we’re giving Google what it wants. We can also use AI to determine search intent, understand the ideal word count for an article, and even plan the structure. All of these things matter.
When it comes to beating those who rely solely on AI-generated content, though, we need to create content that is useful and relatable. At the end of the day, if the content that we’re writing satisfies search intent so that the user doesn’t need to go back to the SERP and click on other search results - if they find the answer they’re looking for in your content - then you always have a better chance. Not just at ranking for one keyword, but at ranking for multiple keywords and gaining legitimate authority that lasts, and authority that drives results.
Machines don’t have empathy, they can’t see the bigger picture, and they don’t have that human element that gives them the bird’s-eye-view we can see. That is the biggest advantage that we’re going to have when we are creating content to beat those who are relying on AI in 2023.”
Should SEOs be using AI just to generate content ideas and concepts, rather than to create content in its entirety?
“Yes. We can use AI to find content ideas and automate the manual process of trying to understand search intent, trying to understand how long content needs to be, and trying to understand what the most important questions we should be answering for a specific keyword are. AI can help us get all of those answers very quickly.
However, we should not be using AI to write content. We can use it to create a brief and to understand what should go into the brief, but not for actually writing the content.”
Will AI miss keyword opportunities to optimise for? Can you trust AI to generate that for you?
“While AI tools can help you to find the most important questions that you should be asking, it doesn’t give the full picture that you get by manually opening SERP results. You need to look at what each person is covering, and especially what they are not covering - because that’s always what you want to be doing better.
What are they missing, that you can exploit and do better? I don’t think you can get that from AI because it just analyses what already exists, and tells you what already exists so that you can do more of the same. AI doesn’t help you find your ‘X factor’. You can only do that by manually analysing the SERP. That’s a strategy that humans will always be better at than AI.
That’s why I say that AI can support you. You can automate determining word count, understanding search intent, and understanding the most important questions, but, when you want to create holistic content, you still have to look at those content pieces that are ranking - to understand what is missing and what could be better.”
Will AI also be replacing podcasts, videos, and content like that anytime soon?
“There is already AI for videos. If you go on AppSumo you will find lots of deals for tools that will create videos for you. You just need to put in a keyword and maybe some ideas about your brand, and it will make the videos using vectors and that kind of thing. They already exist, but is it good? Is it engaging? Does it make you want to take action? No. I haven’t seen any of those videos that made me feel anything. That is why the human element is still better.
For podcasts, AI can potentially help you plan the structure of what you should be talking about - that’s something that AI could be doing in the near future - but for speaking one-to-one, I don’t see that happening.”
Can you beat Google and make it think that AI-generated content has been created by a human?
“People already do that. I used to work in a content mill and that was practically half of our job, you just need to make sure that the article is up to snuff Lens-wise. You make sure to throw in a few infographics (that can be auto-generated with some tools), show a few pictures, some image optimisation, and make sure that it’s using as many entities as possible.
That was always the most important thing that we used to do. We would try to use as many entities as possible because, when the search bots are crawling the page, those are the words that they are looking for to help them understand what that page is about. Our goal was always to try and get those pages to have as many entities as possible, sacrificing quality, and then building backlinks to those pages and linking internally to all the other pieces we’ve written. Most of the time, the clients will still rank.
At the end of the day, Google is a machine and machines can be tricked. I don’t care what Google says, I don’t care how many algorithm updates they release: machines can be tricked. I build content briefs as part of my job, and I analyse at least 20 pieces a day. I see a lot of thin content on SERP and a lot of content pieces that were clearly written by AI. Across all of them, I see the same thing: they are doing this by building backlinks, by populating their websites with a lot of content, and by making sure that they’re optimising for semantic entities.”
Is Jarvis, now called Jasper, one of the better tools? Are there other tools that you’d recommend?
“Jarvis has to be at the top of the food chain when it comes to AI tools for content, but I really like what NeuralText is doing. It’s scary.
When you create content and you want to promote it on social media, NeuralText can help you write all of that content, and it will produce 5, 6, 8, 10, etc., social media posts. It can help you write meta descriptions, intros, ad copy, and a tonne of other stuff that I have not seen any tool out there doing with AI. I started using it recently and it just has so many options for how to automate copy, it’s insane.
The path for writing meta descriptions is very useful, and the path for social media copy can be very useful as well. I don’t write ad copy, but I also believe that it could be helpful to give you the rough idea that you need for inspiration so you can write better copy. It can give you the first couple of lines, maybe. It’s the same for writing a blog introduction or writing the first couple of lines of a subheading.”
Most people seem to be thinking it’s okay to use AI-generated content on descriptions of products or less trafficked pages, but they wouldn’t use AI-generated content on high-traffic pages, blog content, or home pages. Do you agree with that perspective?
“Partly. You can use an AI generator to write meta descriptions because they’re not a ranking factor. You can use it to write social media copy because that’s not going on your website. However, for any content that is going on your website - I don’t care if it’s a lower traffic page - I don’t want to see AI-generated content there.
I don’t want to see AI-generated content there because you don’t know what the next algorithm update is going to be. I always want to be on the good side, so that I’m not panicking when there’s an algorithm update. The best way to do that is to make sure that every single piece of content on your website is written by a human, and is written for humans first and search engines second.
The machines focus 100% on writing for SEO. What we keep seeing with every update, is that Google tries to force SEOs to write for humans first and search engines second. That is the complete opposite of what AI tools do.”
What’s something that SEOs shouldn’t be doing in 2023? What’s seductive in terms of time, but ultimately counterproductive?
“As I’ve already said: don’t use AI to write content. I know there’s a lot of pressure to create as many clusters as possible and get those clusters to run quickly - get the content out quickly - because SEO is always under pressure to generate results fast. Nobody wants to wait three months or six months, and the fastest way to get results is always going to be to create more content.
I understand that pressure but do not give in to the temptation to use AI to write content. At some point, it’s going to bite you in the a**, and you can quote me on that.”
Chima Mmeje is Content Strategist at Zenith Copy and you can find her over at zenithcopy.com.