A year later, research reveals Google’s AI Overviews ‘reimagined’ Search
What you need to know
- A research report by BrightEdge states Google’s AI Overviews have led to a 49% increase in search impressions.
- The post states users are clicking-through less, likely due to AI Overviews appearing in roughly 11% of Search queries.
- Google’s AI Overviews storied history officially began in May 2024, which kicked off its citations update and its trouble with erroneous answers.
Research was conducted on the effects Google’s AI Overviews have on its Search ecosystem, and the results appear positive.
In a report by BrightEdge (via Global Newswire), its research reportedly concludes that Google’s AI Overviews are mostly positive for it and the user experience. The founder and CEO of BrightEdge, Jim Yu, said, “The assumption was that AI would kill traditional search.” Instead, these bite-sized AI notes have increased search impressions “by over 49%,” to which Yu adds, “our data tells a different story: search is not disappearing, it’s expanding. Google is not being replaced; it’s being reimagined.:
The report breaks things down, stating AI Overviews now appear in “over 11%” of Google Search queries. Moreover, the company finds that users are clicking through results less because of this, with a ~30% decline in such behavior.
Citing sources is another topic Google has pushed with its AI Overviews, but it’s not the only one doing this. As the post highlights, other companies like OpenAI with ChatGPT and Perplexity also deliver an “overview” variant. However, as it notes, each is different, with ChatGPT offering “rare” instances of citations. Perplexity is stated to step this up to roughly five citations per answer, yet it mostly skips out on brand names.
BrightEdge analysts claim Google sits between its competitors, offering a mix of brand names and citations that (hopefully) make sense.
An Overview of Google’s AI Overviews
Google tested links in AI Overviews a couple of different times, as the software is prone to either hallucinating or flat-out getting things wrong. With in-line links, users can see clearly (and faster) where the AI got the information it’s reciting. It also helps to retain the AI’s credibility and gives users a chance to ensure they’re not going around spouting nonsense either. Fact-checking has even hit the likes of Gemini.
However, AI Overviews didn’t have a flawless first year, as you may remember the whole “glue cheese on pizza” thing. Reports piled in last spring with users showcasing the strange responses Google’s AI Overviews provided. The issue prompted the company’s Head of Search, Liz Reid, to speak out about them and the cause: “queries that people don’t commonly do”
Reid explained that when there isn’t enough information for the AI to utilize, it creates an information gap. The AI scrambles and often digs up whatever it can to fulfill the query. Unfortunately, the AI dug into troll posts on the internet that were wrong, but it doesn’t know that. Reid stated last year that Google would return to the kitchen to alter how the AI interprets “nonsensical” and “satirical” queries.
Let’s not forget that the AI is drawing knowledge from us, so stop trolling.
After officially launching in May 2024 during I/O 2024, Google has continued to refine and expand its overviews. Most recently, in January, AI Overviews expanded to Circle to Search for the Galaxy S25 series. These overviews include one-tap actions for users who find phone numbers, emails, and more.
Post Comment