When Google CEO Sundar Pichai introduced a generative AI feature for the company's search engine last month, he and his colleagues demonstrated the new capability with six text-based queries for the public to try .
Questions included “how do I clean a fabric sofa” and “what should I use to get a coffee stain out of my carpet.” These were intended to highlight how Google's new feature, AI Overviews, could generate comprehensive and useful summaries of information on top of traditional search results.
But on Friday, only one of six queries still provided an overview of the AI, according to The New York Times' tests. Instead, the functionality was noticeably less widespread. A search for “what should I use to get a coffee stain out of my carpet” now resulted in a snippet of text from a website, JDog Carpet Cleaning & Floor Care, while “how to clean a fabric sofa” was replaced by a link to the HGTV website with the answer. (Search results may vary by user and location.)
The disappearance of AI overviews for some searches appeared to be part of a broader step back after the new technology produced a litany of falsehoods and errors, including recommending glue as part of a pizza recipe and suggesting people to ingest rocks to obtain nutrients. Users complained loudly on social media about the errors, in many cases openly mocking Google.
Liz Reid, Google's recently promoted head of search, wrote in a blog post Thursday that the company has pared down its AI overviews in certain ways, rolling out “additional opt-in refinements” to deliver more health-conscious answers , disabling misleading advice and limiting the inclusion of satire and user responses from forums like Reddit.
“We will continue to improve when and how we show AI overviews and strengthen our protections,” he wrote, adding that Google is working on updates to improve broad sets of search results.
Ashley Thompson, a spokeswoman for Google, said in a statement Friday that the company has made more than a dozen technical updates to its systems.
“AI overviews help people answer a large number of Search queries today, serving as a starting point for content across the web.” The company added that while it was making changes to improve AI overviews, it would not be backing away from the feature long-term.
The step back was a blow to Google's efforts to keep pace with its rivals Microsoft and OpenAI, the maker of the ChatGPT chatbot, in the frantic race to lead in artificial intelligence. He also highlighted the difficult strategic choice Google faces over whether to embrace AI technology that may not be reliable or keep the same hugely popular search engine and risk falling behind its peers.
Google had chosen to proceed more slowly than Microsoft, which put more conversational AI into its Bing search engine early last year. Google, which has many more users than Bing, tested AI features for its search engine a year before introducing AI overviews. The company said the new feature will roll out immediately to users in the United States and to more than a billion people by the end of the year.
But ultimately, Google “should have rolled it out more slowly,” said Patrick Hall, assistant professor of decision sciences at the George Washington University School of Business. “When something like this happens, you really have to retreat. And if nothing else there is reputational damage” for the company.
Google, which has led internet search for more than two decades, has struggled since OpenAI released ChatGPT in 2022. Some tech industry insiders have considered the chatbot's ability to generate responses a serious threat to the search engine. Google search, which has been the most popular method for obtaining information online.
Since then, Google has worked aggressively to regain its lead in artificial intelligence, releasing a family of technologies called Gemini, including new AI models for developers. The company has also integrated the technology into YouTube, Gmail and Docs, helping users create videos, emails and drafts with less effort.
Last month, Ms Reid said at Google's developers conference that the search engine would do more Google searches for users with AI Overviews. You highlighted increasingly complex requests that Google could address with the feature, but those features have not yet rolled out to users.
On stage, simpler questions like “how do I get the smell of a campfire out of my clothes” produced AI-based answers, including airing out, adding baking soda and sprinkling with lemon juice.
But when users got their hands on the new service, they found that the AI overviews sometimes generated incorrect, or even dangerous, responses, including the recommendation to put non-toxic glue on the pizza to make the cheese stick. He also misunderstood some cited websites and misinterpreted the presidential story.
Of the six questions Google threw out to the public, the only one that consistently triggered an AI overview on Friday was “what are some interesting science projects I can do with my 12-year-old son.”
The answer? Grow crystals, extract DNA from saliva and write a message with invisible ink.