Alrich Lawson | Getty Images
Last month, we looked into some of the most inaccurate, dangerous, and downright bizarre answers generated by Google's new AI Summary feature. Since then, Google has reportedly issued a partial apology/explanation for generating such results and reversed the rollout of the feature for at least some types of queries.
But the more I think about this development, the more I question the wisdom of Google's AI-powered search results in the first place. Even if the system didn't blatantly erroneously produce results, boiling down search results to simple, concise, AI-generated summaries seems like a fundamental misunderstanding of how people use Google in the first place.
Credibility and relevance
When people type a question into the Google search bar, they're rarely looking for basic reference information that can be found on a Wikipedia page or a company's website (or even in a Google snippet of information). More often than not, they're looking for subjective information for which there is no single “right” answer, such as “What's the best Mexican restaurant in Santa Fe?”, “What to do with kids on a rainy day?”, or “How do you stop the cheese from sliding off pizza?”
Google's value has always been in taking you where Google thinks. probably Getting good answers to these questions can be difficult, but it is up to you, the user, to determine which sources are the most reliable and relevant to the information you need at that time.
-
The guys at Pep Boys weren't amused when they said this either.via)
Kyle Orlando / Google
-
Weird Al also recommends “running with scissors”!via)
Kyle Orlando / Google
-
This list of steps is actually taken from a response in a forum thread about how to do something entirely different: (via)
Kyle Orlando / Google
-
An island that's part of the mainland? (via)
Kyle Orlando / Google
-
Everything is cheaper now, so why does everything seem so expensive?
Kyle Orlando / Google
-
This Truman should not have been president… (via)
Kyle Orlando / Google
When it comes to credibility, any Internet-savvy user will utilize a myriad of contextual clues when judging a random Internet search result: Do I recognize the medium or author? Does the information come from someone who appears to have expertise or professional experience, or is it just a random forum poster? Is the site well designed? Has it been around for a long time? Does it cite other reliable sources? Etc.
But Google can't know in advance which results will fit what you're looking for. For example, when it comes to restaurants in Santa Fe, do you want an authoritative list from a respected newspaper critic, or quirky suggestions from a local? Or maybe you scroll down a bit and stumble upon a less relevant article about the history of Mexican cuisine's influence in the city.
One of the intangible strengths of Google's search algorithm is that it lets users decide which results are best for them: if the first few pages of results contain reliable and relevant information, it doesn't matter if the other links are “wrong” for that search or for that user.
