Talk of artificial intelligence and machine learning have hit a fever pitch in the last few years as the technology behind them have advanced at a rapid pace. However, as far as they’ve come, they still have a long way to go.
I like to compare the behavior of AI to that of a drunk toddler. Sometimes it gets things right, but just as often, it gets things wrong, and every so often, it does something completely embarrassing, ridiculous, and even terrifying.
Artificial intelligence, powered by machine learning, has the potential to be an amazing and positive force in powering search engines, but it still has a lot of growing up to do.
Right now, in its infancy, it’s still crawling around in Garanimals PJs with a pacifier in its mouth. In a few years, it will be more like an angsty teenager screaming “You don’t understand me!” while eating Tide Pods — but it will also be a lot closer to delivering useful results more consistently.
And by results, I don’t just mean spitting out 10 choices, or even one, in response to a query plugged into a search field. I mean that it will go so far outside of the traditional concept of a search engine by anticipating and predicting our needs long before we even think about them.
I’ll call this predictive search.
Here’s an example: I can sometimes be geographically challenged, so when I leave an appointment, I typically open my map app and click my office address to guide me back.
This is a perfect example where AI and machine learning can shine.
When my appointment is over, rather than me having to manually open my map app and enter the address of my next appointment, Google can evaluate whether I’ve been to the address of my next appointment before, and spring into action by either:
Automatically opening my map app with the next address on my calendar ready to go if it’s somewhere new,
Offering a prompt if it’s somewhere I’ve been a few times,
Or simply not opening at all if it’s somewhere I go frequently.