Click here to talk to an eCommerce SEO Expert!

AI really has redefined what it means to be a buzzword. Seldom if ever before have I heard similar terms thrown around like confetti, with abject and reckless abandon.

For some in digital marketing, specifically SEO, AI has been touted as the savior. In some ways, content is the most important pillar of SEO, and it certainly is the pillar on which authority and credibility both firmly rest. 

AI has promised to automate the entire content process, from keyword research to copy generation and optimization. It has promised to produce better outputs and cut costs across the board.

Except it hasn’t. Well, it may cut costs, but it doesn’t produce better outputs. In fact, it produces far, far worse outputs than I even thought were possible. 

I wouldn’t believe some of these things if I hadn’t seen them firsthand.  

The Broader Problem

Before I launch completely into my full-bore anti-AI spiel, allow me to frame the issue with some real world facts. 

Feast your eyes on the following article about the use of AI in litigation. Yes. It really happened. Real, actual, barred lawyers asked ChatGPT to prepare a brief for a legal proceeding. But it gets better. ChatGPT made up legal precedents. And then a real, actual lawyer who went to law school presented them in court before a judge. 

Dear reader, it did not go well. 

This is only one instance of many. There are so many of them I could actually probably write an entire article on the cringe-inducing, jaw-dropping things that AI models are churning out for people – even when presented with reasonable prompts. 

Here are some a few of them: 

And let’s not forget when Bard publicly, before perhaps thousands of people, erroneously claimed that the first picture of an exoplanet was taken by the James Webb Space Telescope. (It wasn’t.)

Forgivable, no?

Not to me, and not to any who seek the truth, which appears to be an increasingly difficult thing to find nowadays. 

Which is actually my first reason that you should not use AI in SEO.

A cool stock photo which appeals to emotion and forgets the fact that AI gets the answers to simple questions stupidly wrong.

Reason 1: L-AI to Me 

The reason content is such an important pillar of SEO is because content contains information. This is the crux of SEO, the foundation on which the entire edifice rests. If the content be corrupt, the whole structure be a poison tree.

Err, I mean, if you have false information on your website, your SEO score will slide over time. 

Now, as I have demonstrated already, AI is critically bad at presenting information and often gets it wrong. In some instances it apparently blatantly makes things up. 

But I have better examples. I literally just asked one of the free versions of AI for some good winter tactics for fishing for bass. 

Admittedly, it appears to have improved. It suggests that I fish deep and very slowly. Check and check. It has the potential to deliver good information.

But here’s where things go haywire. It also suggested I use a small bait. Now, downsizing your baits in the winter is a recognized tactic for fishing, but failing to mention that lethargic fish are more willing to expend energy to strike a bait if they can justify that expense through a larger meal is egregious bordering on disingenuous. 

That is to say, ChatGPT didn’t tell me to use larger baits, which is a much more widely acknowledged winter fishing technique than downsizing. 

I recall in the past doing research for a client wherein I asked ChatGPT for tips for hunting pheasants without a dog. What I got back was comedy, but not the kind that makes you laugh. 

It told me to use scent control and a decoy. It also told me to be very quiet so as not to spook the birds. My friends, I hope some of you reading this are hunters so you can realize how fatally bad this “advice” is.

Of course these are not things you would not know if you were not an angler or a hunter, and that’s where the problem lies. 

Now, the danger here is not when AI does not know better. It’s when the person asking for something doesn’t know better. 

You might ask ChatGPT something, get an output, and if you take it as gospel, you will at best be garnering false information and at worst putting yourself in peril.

Why else would OpenAI have a whole page dedicated to safety?

What is the danger here, people? 

I digress. The point is that the biggest argument not to use AI in SEO is that it gets it wrong. Unfortunate, but true. Until it gets much better, you simply can’t use it to perform reliable research. 

Reason 2: It’s Theft 

AI is subject to a condition called data bias, which means it can only produce outputs based on the data it was trained with. This has a whole load of implications (including some related to ethics). 

It also harkens back to the old adage, “garbage in, garbage out.” The long and short of what I’m getting at here is AI is limited in what it can put out by what is put in. 

And that means sometimes AI straight up steals copy from elsewhere, which can get you a duplicate penalty. 

I wouldn’t gamble with duplicate copy or plagiarized copy. It might not get you an outright penalty from Google, but it will not build your rankings.

And besides, it will take you longer to fact-check AI and vet it for plagiarism than it would take if you just did the research and drafted copy the old way. You know, the way that works?

Reason 3: You Cannot Build Authority with AI 

Remember E-A-T? It’s actually E-E-A-T now. That’s Expertise-Experience-Authority-Trust. 

Now, technically, if you vet the AI outputs you use on your website, you could build trust, if the information is accurate. 

But authority does not just come from accurate information. It comes from experience – and AI cannot produce an experiential component. 

Forget the fact that you need a human author for the “expertise” part. I’m not sure if Google is awarding rankings by cross-referencing the articles and posts published by the same author on different websites or even within one domain. 

But here’s what I can say. Authority comes from that experience, and it is sorely lacking in AI. Human writers, regardless of the topic they are writing on, can capitalize on their own experiences, reflecting, extrapolating, and applying reason to those experiences in order to present digestible information, sometimes containing actionable advice, to whatever readers come their way. 

AI can only produce based on the data on which it is trained. This is, by its nature, limited. Theoretically the entire data set of the whole, extant human experience is limited, too – but guess what, it is way bigger than whatever data set was used to train any given AI model. 

And that means you’re getting a way broader pool of experience when you lean on actual writers rather than AI.

Who knows, AI can improve. I certainly won’t say it can’t get better, and perhaps with further iterations it will actually become better than any one given human contributor. But that time is most certainly not now.

Reason 4: It’s Cringy and Spammy Beyond Your Wildest Dreams 

AI output is so bad, you guys. I’m not going to overdo this section. If you don’t believe me, visit OpenAI and use one of their tools. 

Ask it a question, then ask it a similar question and tell me it doesn’t repackage the first output it originally gave you. 

It is repetitive, flat, and spammy. It seems novel for an iteration or two but then that quickly goes down the drain. 

By the way, Google does not like spam.

Reason 5: It’s Categorical Black Hat Nonsense 

Remember how SEO “experts” used to boost rankings with spam content, keyword stuffing, and copy hidden in page margins or other “no-access” portions of pages? 

And then do you remember how the domains that invested in those strategies basically collapsed and were unable to recover? 

Yeah, that was called black hat SEO. Today, those techniques don’t work because Google recognizes them, but it doesn’t mean black hat SEO is dead. 

It’s just a different animal, and one of those new “black hat” techniques is to use AI. 

If you don’t believe me, just check this out

Straight from the horse’s mouth. Google will crack down on you if you do it. Actually, in my line of work I’ve already seen this happen. 

Don’t say you haven’t been warned. 

Reason 6: It Flops at Technical SEO 

Apparently AI flops hard at coding. I’m not 100% sure about this because I’m not a developer, it’s just the vibe I’m getting from people that do know. 

That means AI is also categorically bad at technical SEO, which revolves around structured data, site structure, streamlined integration, and issues with security and speed. 

Would you relinquish technical command of your website to a tool that can’t get basic questions right? I know I wouldn’t.

Anyway, do your own research here. Though I can speak authoritatively on the other aspects of AI’s shortcomings, this one is little more than a rumor to me.

But, considering the other execrable qualities of the current AI landscape, I think it merits further investigation. 

Reason 7: AI Cannot (or Simply Does Not) Keep Up with Algorithm Updates

Finally, AI either cannot or does not keep up with Google’s algorithm updates, unlike an actual SEO expert that reads the news and keeps abreast of updates and other best practices and trends. 

Also, for instance, ChatGPT was trained on data the most recent of which is from 2021, which means if you ask it for current stats or data, you’re going to be working with outdated information. 

Not a big deal across the board, but still something to consider. Anyway, you don’t want to lean too heavily on AI only to find out a few months or even a year from now that it was a bad idea all along. 

It’s not like you didn’t know better, right?

It takes longer to do SEO the old-fashioned way, but nothing worth having comes easily.

The Solution: Work with Humans 

Before you charge and crucify me for what you might call protectionist, isolationist rhetoric, be apprised that I would like to intercept that charge on the grounds of self awareness. 

I am fully cognizant of my position as a copywriter and that, conceivably, if AI could do my job as well as I can (it can’t) I could be cost-effectively replaced. 

This is not unknown to me. But the truth is simply that. I am telling the truth. You should not use AI in SEO for all of the reasons I have enumerated here, and perhaps others. 

It’s not like I don’t have other skills. Believe me, I love writing, but I am in no way attempting to protect my industry by publishing biased, dishonest drivel that paints these machines like the precursor to Automated Hell. 

I’m just sharing my experience. I know firsthand what authoritative copy and SEO best practices (executed by thinking individuals) can do. And I also know firsthand what serious low quality results AI generates. 

So, make your choice. I’ve done my part.

Like
Share

Michael Esposito

Mike Esposito is a professional SEO copywriter spurned by a love of language and creativity. When he's not at the keyboard, you may be able to catch a rare glimpse of him enjoying the outdoors or sipping fine literature.

Read All Articles

Leave a Reply

Your email address will not be published. Required fields are marked *

Get a Free SEO Audit

Get a Free SEO Audit