AI Overviews in Google Search

Google Search is getting a makeover, and it's called AI Overviews. This new feature promises to understand complex queries and provide summaries of the answer, pulling information from various sources. It sounds great in theory - one stop for all your search needs. But is it too good to be true?

AI Overviews: A TL;DR (too long-didn’t read) for Your Search
Imagine searching for a yoga studio. In the past, you might have searched for "yoga studios near me," "popular yoga studios," and "yoga studio discounts." Now, with AI Overviews, you can ask for all three at once. The overview will summarise the info, potentially including popular studios in your area, their commute times, and any discounts they offer.
This "generative AI" aims to make search results more visual and interactive. Think images, videos, and a more social search experience.
Are the Answers Actually Right?
Here's the catch: AI Overviews are making some big mistakes. While it can answer basic questions accurately, reports show it has provided incorrect information about everything from running with scissors to dogs playing on the oval during a footy match.
During the Google I/O presentation, there were some absolute clunkers - right out there warts and all for the world to see first hand that Google has a way to go.

1. How to keep cheese on pizza? Add an eighth of a cup of nontoxic glue. That's a tip that originated from an 11-year-old comment on Reddit.
   

1. In response to a query about daily rock intake for a person, it recommended we eat "at least one small rock per day." That advice, from a story featured in TheOnion.com in 2021.
   

How AI Gets Confused
So, how does an AI designed to help end up suggesting rock consumption? AI hallucinations stem from a few factors.

• Flawed Training Data: AI models learn from the data they're fed. If that data includes inaccurate or misleading information, the AI can incorporate those errors into its responses.
  
• Algorithmic Errors: The complex algorithms that power AI can sometimes malfunction, leading to nonsensical outputs.
  
• Context Misunderstandings: AI might struggle to grasp the nuances of human language and misinterpret the intent behind a query.
  

It’s basically a "Garbage In, Garbage Out" Problem
Let’s hope Google can take out the garbage, before search as we’ve come to know it - becomes just that!
AI Overview - coming to Google Search soon 🙂

​