Have you ever searched for something on Google (or another search engine of your choice) and had to scroll through the results, clicking one link after another, trying to find the answer you were looking for?
Well, now Google is confident that their AI system can “do the searching for you.”
Wasn’t their algorithm already doing the searching for us? Hasn’t that been Google’s whole thing for the past 26 years?
Table of Contents
How AI Answers Work in Google
The difference is that Google now thinks it can use “generative” AI to give you the answers directly so you don’t have to click on links to get the answer you need.
What that means is you still see a bunch of links you can choose to click through for more information, but you’ll also see a section of text that was “generated” by AI.
I use “generated” in quotes because AI doesn’t truly generate anything. It just scrapes the internet for existing content on related topics, then repurposes it into something it thinks you want to read.
The problem with this is that AI is either unable to filter certain pieces of content based on context, or just chooses to ignore the context and assume that if the content was found on a related piece of content, it must be relevant to the question being asked.
This is how someone ended up with Google’s AI section telling them to use glue to prevent the cheese from sliding off their pizza. Apparently, the “advice” was taken from a Reddit post that was obviously meant to be a joke, but because AI ignored (or was unable to account for) the context, it provided this content as a valid response to a question about pizza.
The other problem we’ve been having with AI is that it doesn’t have the same knowledge base we have. As humans, we know we can’t (or at least shouldn’t) eat glue. It’s so obvious to us that we haven’t thought to teach computers that humans can’t eat glue. This is part of what has led to some of these hilarious mistakes “generative” AI has been making.
While the glue joke is funny, it’s important to remember that AI can just as easily mess up in ways that could be much more damaging if it’s not obvious to us that the information is wrong. Consider all the dangerous diet advice out there that real people are posting as real solutions to help people lose weight. If AI doesn’t know what we can and cannot eat, it certainly can’t tell the difference between a healthy way for us to lose weight and a dangerous way for us to lose weight.
To be clear, there is a disclaimer at the bottom of the AI answer section reminding us that this is still in beta form and Google is working to correct the gaps in information that have led to their AI system giving hilariously wrong answers, but I doubt they’ll ever be able to completely rid themselves of the problem.
AI Answers vs Featured Snippets
Let’s not forget that Google had been heading in this direction for a couple of years already. Any time you asked Google a question that had a short answer, it would provide the answer to your question directly on the Google result page, so you wouldn’t have to click through to the website to get the answer.
Despite the fact that Google has been making a really big deal about this announcement, I’m not sure I see a significant difference between this and their previous zero-click search results. The results you get in the “AI overview” are slightly longer than the ones you previously got from Google’s zero-click search results, but not significantly longer.
The main difference is that Google’s previous zero-click search results would be pulled from one page on one website, which you could then click through to get more information if that’s what you wanted to do.
The new AI overview is not only longer, but it’s pulling information from multiple sources. It then provides you with the sources it’s using to provide you with that information so you can click on one or all of them to get more information.
The thing I like about that setup is that it enables you to verify the information you’re getting from Google’s AI by seeing where that information is coming from. That way, when Google references a joke and tries to present it as a valid answer to a question, you can see where that “information” came from.
Why Is Google Doing This?
When you stop to think about it, it’s not really surprising that Google would incorporate this type of capability into their search engine, given the fact that people started using ChatGPT as a search engine almost as soon as it was released to the public. Google needs to compete, and that’s what it’s doing.
Regardless of where this ends up going, I do appreciate the way in which Google is trying to adapt to this brave new world of “generative” AI in which we find ourselves today. Aside from the fact that ChatGPT is not a search engine, my other big problem with it is the fact that it doesn’t give you a chance to click through to the websites where it got its information from, which means you have no idea where that information is coming from. And you have no opportunity to compare answers.
For example, as temperatures climb here in Chicago, I was recently having a conversation with some people about a rumor we’ve all heard that drinking hot beverages when it’s hot out can actually help cool you down. Is this true, or just another myth?
I decided to ask my good friend, Google, but I didn’t take the first answer I found at face value. Knowing about the prevalence of misinformation online, especially when it comes to health and biology, I decided to click through at least a couple more websites before determining whether to believe what I was reading. When I found three different sources that all agreed with each other, I felt pretty confident I had the right answer.
When I was researching how often I should bathe my husky, I found a bunch of answers that all agreed with each other and one outlier. I ignored the outlier, especially since one of the other answers was provided by a veterinarian. I’m definitely going to take her word over some random stranger who may or may not know what they’re talking about.
If you found this content useful and would like to have more like it sent straight to your inbox, you can sign up for our newsletter in the space below. You can also see what we’re up to by following us on LinkedIn, YouTube, and Facebook.
