Google is in damage control mode following AI Overviews fiasco

Summary

• Google’s AI Overviews have made the headlines recently for generating inaccurate and sometimes dangerous results.
• The company has now defended AI Overviews in a statement, suggesting some viral posts highlighting the issue may have been doctored.
• Google announced this month that AI Overviews would be the default way to search in the US, with other countries following suit imminently.

Google this month made AI Overviews the default for search results in the US, with plans to bring it to other parts of the world soon. However, the past several days have been filled with examples of these AI-generated results getting things wrong or providing dangerous recommendations. To say it has been a tough couple of weeks for the AI Overviews team would be putting it mildly. In the midst of a social media storm over this change, Google has now put out a statement dismissing some of these viral posts while promising to take action against the wrong results.

In a statement distributed to multiple media outlets, the search giant said that AI Overviews provide “high-quality information” on most occasions (via 9to5Google). Furthermore, Google said it couldn’t replicate some of the viral AI Overviews, while also reportedly encountering “examples that were doctored.” Here’s Google’s statement in full:

The vast majority of AI Overviews provide high quality information, with links to dig deeper on the web. Many of the examples we’ve seen have been uncommon queries, and we’ve also seen examples that were doctored or that we couldn’t reproduce. We conducted extensive testing before launching this new experience, and as with other features we’ve launched in Search, we appreciate the feedback. We’re taking swift action where appropriate under our content policies, and using these examples to develop broader improvements to our systems, some of which have already started to roll out.

The mention of “taking swift action” refers to removing some of these inaccurate results, like suggesting that astronauts have met cats on the moon, as an Associated Press reporter found out. Meanwhile, some were also recommended to put glue in pizzas as a way to get the cheese to stick. Thankfully, these incorrect and hazardous overviews have since been removed, with Google potentially having to keep a closer eye on some of these slip-ups in the future.

There are ways to bypass AI Overviews

In response to the initial uproar about AI Overviews getting things wrong, Google rolled out a Web filter for search that would only provide links related to the query, without any of the AI stuff. It was later discovered that adding &udm=14 to the end of a Google Search URL would take users to the links-only results page. There’s also a fun website created by Ernie Smith — udm14.com — which offers a seamless Google Search experience without any of the clutter.

The controversy surrounding AI Overviews has some parallels to the Google Gemini image generation debacle not too long ago. However, it doesn’t look like Google is ready to hit pause on AI Overviews just yet, with a recent blog post saying, “People who use AI Overviews actually use Search more and are more satisfied with their results.” The company also confirmed in the same post that it would bring ads to some of these AI-generated results.