Why AI (Probably) Will Not Replace Google

All the digital marketing “gurus” that sound off that “SEO is dead” every New Year are having a field day with the doom signal du jour: AI is going to replace Google. 

Say goodbye to your current marketing strategy, they clamor, because Google search will soon be replaced by ChatGPT or some other AI megamodel. 

Soon (though they’re not sure how soon) they argue, consumers will regard Google search as the cassette tape of the digital age, fondly remembered, but functionless.

Except, that’s not happening. And while I am humble enough (unlike many other digital marketers that shall remain nameless) to admit I lack the gift of foresight, the writing on the wall does not, in fact, suggest that Google search is moribund. 

And these are the reasons why. 

The Sheer Number of People That Use Google

If anything speaks testimony to my claim that AI is not going to replace Google, it’s the sheer number of users that Google has, not to mention its market share. 

Google is getting something like 8.5 billion searches per day, and more than 9 million searches per minute. That’s more than a search per day per person on the planet, and let’s not forget, large portions of the global population aren’t even electrified, let alone connected to the internet. 

By contrast, ChatGPT gets an estimated 2.5 billion searches per day. That’s not nothing, considering, but Google still gets about 3 times that amount. 

You can also drill down and see what portion of the market Google controls. Google owns more than 90% of global search market share. 

Now, these figures could change, but…

They Do Fundamentally Different Things 

There is some overlap in their utility, but by and large, Google and ChatGPT do fundamentally different things. Google is like a very fast encyclopedia, and ChatGPT is more of a digital assistant. One provides general information, the other provides curated information. 

Now, these two models are valuable for very, very different reasons. If I want to know the answer to a very specific question, very quickly, ChatGPT is better than Google. But if I want to get educated, broadly, on a topic, ChatGPT is not just bad, but very bad. Google, or another search engine (or dare I say, a book) is much better. 

That doesn’t even address some of the issues with AI models, like hallucination (more on that in a minute).

What Google Is Good at, AI Isn’t (Yet) 

One of the great things about search engines (not just Google) is the manner in which it furnishes information. It’s not as selectively curated as AI models like ChatGPT. 

The regurgitated output of ChatGPT may be convenient but it is not helpful for getting a holistic view of what you’re looking for. Sometimes the search results are fairly nonspecific, and in an interesting way, sometimes that’s very helpful. 

Utilizing only the highly-specified outputs of AI models is like putting blinders on. Good if you want to know what the temperature is in your town right now, but not very good if you want to know why the prevailing weather patterns are in your area of the country. 

There Are Limitations to AI Right Now 

AI discourages active thought. That, coupled with the fact that AI gets plain answers to pretty straightforward questions patently and abjectly wrong (see the immediately following section) is reason enough that AI is too limited to replace Google. 

Even if it were not, AI doesn’t offer a lot of the features and applications that Google has. For instance, there is currently no AI analog to Google maps, which makes AI useless for navigation or finding local businesses. AI only recently adapted to eCommerce; previously, you couldn’t buy through it, which also severely limited its real-world applications, as well its ability to be used for profit. 

Also, even if AI eCommerce does take off, there are some shoppers that just aren’t going to use it. For some, eCommerce has replaced brick and mortar only in the sense of where the transaction occurs. Many online shoppers still like to go through all the proverbial (albeit digital) aisles before making a purchase. Shoppers like these don’t want to be spoon fed things to buy. They want to find them. It’s not shopping otherwise. 

AI Hallucinates 

This is perhaps the biggest issue with AI right now (I know, that’s a bold claim on my part). Google can serve you bad information, but the thing is, Google isn’t the author. It’s just trying to connect you with information. Whether it serves you accurate or inaccurate information, the onus is on you as a researcher, or student, to figure that out for yourself, you know, by thinking. 

I cannot tell you the number of times AI models have given me laughably, even dangerously bad, patently wrong information, but it has happened enough times that it has effectively put me off using them for good. I just don’t see the use case. 

The way I see it, when I want to learn something, I still have certain responsibilities as a student, and I will never relinquish the right to think freely. 

Also, this is not a permanent road block to AI’s projected takeover of the search industry, because if we can solve the problem of AI hallucinations, then it won’t be an issue anymore.

AI Needs Google (or, the Internet, Rather) 

As good as AI is at what it does, until things change fundamentally, it’s dead in the water without the internet, more specifically, without Google. 

Getting indexed and ranking on Google is the first step to showing up as an AI output. Think of Google like the encyclopedia of information and answer engines as just a quicker, more specific model. 

They take what’s already indexed and ranking in Google and re-synthesize the information in a more succinct, digestible format according to whatever prompt you enter into them. 

Until that changes, Google is not going anywhere. AI is just, as it seems, another way to use search engines after all. 

Why SEO Is Still King 

Lastly, I think it’s important to draw attention to the fact that there are still tons of people that use search engines for their search for information rather than AI models. 

That, paired with the fact that search engines do many things better than AI models, as I’ve demonstrated, and that answer engines need search engines in the first place, makes a pretty compelling case for their continued viability. 

That being said, I feel the need to disclaim this because anything could change – but the way it stands, AI-powered answer engines just aren’t going to replace Google, or make search engines obsolete. 

The Likely Future (That We Already See): Integration 

Alright, so we’ve arrived at the conclusion that AI is not going to replace Google. What, then, is going to happen? 

This is not a matter of speculation. We can already see the answer in the search results through Google’s Search Generative Experience. 

AI is not going to replace search engines, but they are going to integrate it into the experience they produce to stay flexible and deliver on user expectations. 

This can already be seen through Google’s AI overview, which has fundamentally changed how users interact with the search results. Already it has given rise to zero-click search. 

So, AI is not going to replace Google, but it is going to change how people use it. It already has.