Google makes fixes to AI-generated search summaries after outlandish answers went viral

abcnews.go.com:

Google said Friday it has made “more than a dozen technical improvements” to its artificial intelligence systems after its retooled search engine was found spitting out erroneous information

Google said Friday it has made “more than a dozen technical improvements” to its artificial intelligence systems after its retooled search engine was found spitting out erroneous information.

The tech company unleashed a makeover of its search engine in mid-May that frequently provides AI-generated summaries on top of search results. Soon after, social media users began sharing screenshots of its most outlandish answers.

Google has largely defended its AI overviews feature, saying it is typically accurate and was tested extensively beforehand. But Liz Reid, the head of Google's search business, acknowledged in a blog post Friday that “some odd, inaccurate or unhelpful AI Overviews certainly did show up.”

While many of the examples were silly, others were dangerous or harmful falsehoods. Adding to the furor, some people also made faked screenshots purporting to show even more ridiculous answers that Google never generated. A few of those fakes were also widely shared on social media.

The Associated Press last week asked Google about which wild mushrooms to eat, and it responded with a lengthy AI-generated summary that was mostly technically correct, but “a lot of information is missing that could have the potential to be sickening or even fatal,” said Mary Catherine Aime, a professor of mycology and botany at Purdue University who reviewed Google's response to the AP's query.

For example, information about mushrooms known as puffballs was “more or less correct,” she said, but Google's overview emphasized looking for those with