The use of artificial intelligence in search - article by David Amerland on

“Anyone who has ever watched an episode of Star Trek will have noticed that on the USS Starship Enterprise computing is both ubiquitous and intelligent. There’s never a need to manually input any data in the Star Trek computer and the only way to interface with it is with voice commands. The captain speaks to it the same way he’d speak to an individual.
“If that was not wonderous enough the Star Trek computer seems to be able to understand spoken speech without any problem (and presumably speaks many languages, not just English) and it collects its own data through its sensors making it therefore both instantly scaleable and independent of its operators.”

That’s the opening of just the second chapter of Google Semantic Search (a free copy of which can be downloaded here). Given the current developments in search where Microsoft has added ChatGPT to Bing and forced Google to respond with Bard, kind of makes the book, first published in 2013, prophetic and its author, (me) - a genius. 

Except, of course, none of that is actually true. Writing it, more than a full year before its publication date I had access to the same cutting-edge research and data I have today and the direction towards which search was rapidly heading was obvious to me. The reason it was obvious to me back then, is the same reason that, today, I am not too worried that Google will become obsolete, that SEO will die or that search will be a thing of the past. 

In writing this piece then (because I don’t often write on search these days) I am giving you the same benefit of insight I possess so you can better decide your actions on marketing, search engine optimization (SEO) and branding. 

How Does Artificial Intelligence Change Search?

The short answer is it doesn’t. Search is the means through which we navigate the web. We do so through some kind of interface where we input a query and the interface provides an answer to that query. That hasn’t changed and it is unlikely to until we get to personalized agent bots that will know us well enough to work pro-actively and predict our needs and get them for us, which is something that was first discussed at the academic research level back in 1999(!). 

We’re not there yet, though this will be in our very near future. Where we’re at, right now, is in the expression of the near “perfect search engine” that Eric Schmidt, Google’s former CEO, described at the Zeitgeist '06 conference in London when he said that "The ultimate search engine would understand everything in the world. It would understand everything that you asked it, and give you back the exact right thing instantly,"

That future is kinda here now. And it isn’t. It’s here in terms of look and feel as ChatGPT becomes the new kid on the block with 100 million active users in just two months. It isn’t in terms of the description Schmidt gave back in 2006. No search engine right now can understand everything that you ask of it and the accuracy in the answers provided still gives a substantial lead to Google that is the company’s to lose. 

Why Is Search Still Inaccurate?

Search hasn’t fundamentally improved in terms of what it does because the landscape it operates in is largely the same. The web is still made of data that needs to be indexed, organized and classified in order to be understood. Data is still governed by the 4Vs of Volume, Velocity, Variety and Veracity. Search queries are still driven by behavior. Behavior is driven by intent. And intent is dependent on context. 

Searcher behavior is dynamic precisely because there is a fluidity in intent and context which changes the value of the underlying data. 

Google’s overall advantage in search was the sheer amount of data it had access to at a personal, global level. This hasn’t changed. It is still the world’s leading search engine with a treasure trove of personal profiles and a strong understanding of the cross-over from desktop to mobile, thanks to its Android operating system, installed in smartphones. 

Is Google Right To Panic?

Google’s rapid integration in search of Bard its AI chatbot based on its large language model (LLM) called LaMDA, which is short for Language Model for Dialogue Applications suggests that the company understands the danger posed by perception. 

Everything that has happened in search to date is an evolution of two tandem forces: 1. How a search engine indexes the web. 2. What people look for when they carry out a search. The first of these is technological. The latter is purely behavioral and it arises out of real needs and the perception of capability. For example, fewer people used search for trivial questions when search needed the input of search operands to deliver more accurate results. The effort of knowing what to use and remembering it was just not worth it. 

When search becomes simple more and more people use it for what appear to be inconsequential things. But that is the entry threshold. The more people use search the more people provide data (which search engines need) and the more people see ads (which is how search engines make a profit). 

Google is worried here not so much about the accuracy of results delivered in search (which is how we should actually grade a search engine) but on the perceived effort required to carry out a search query and understand its answer. Back in 2011 Larry Page said that “All of us at Google want to create services that people across the world use twice a day — just like a toothbrush!

If that changes Google would lose the eyeballs it needs to make the profit it requires from search. So, the company’s anxiety to the perceived threat posed by chatbots in search is fully justified. 

Does That Change SEO or Marketing or Branding?

Of far more value to us is the question of whether searcher behavior should affect the way we do SEO, marketing and branding (all of which are elements that feed into higher visibility in search). The answer to that is a solid “No, but …” 

Obviously the basics have not changed. The fundamentals are still the same. But SEO, marketing and branding are always retroactive. They chase end-user behavior in order to meet expectations and elicit a desired response. As the online/offline divide becomes increasingly blurred, the ability to understand the audience and pivot with it to maintain the connection is key. 

And that, though hardly rocket science, is truly magical because human behavior is not always predictable. How we use the more advanced tools we have at our disposal to become better at seeing it, understanding it and then predicting it is key to business success. 


Go Deeper: 

Take Control Of Your Actions.    Make Better Decisions.