Google is making some changes to its AI overviews after its artificial intelligence-powered search function gave “strange and erroneous” responses to people’s online searches.
AI Overviews were introduced last month at Google’s annual I/O developer conference. Now when people use Google Search to find information about certain topics, a box with AI-generated text appears at the top of the search results, with links to external websites. Traditional search results appear below the AI overviews, marking a major change in the way Google presents information.
According to a blog post by Google VP Liz Reid, AI Overviews results are generated using the company’s large language model (LLM), Gemini, and are designed for cases where someone “wants to get a quick overview of a topic as well as links to learn more.”
Google technology expert Alex Joseph told ABC Audio that AI Overviews can answer more complex questions than a traditional Google search.
“With an AI overview what [Google] what you can really do is synthesize a lot of information and give you the answer you’re looking for very quickly,” Joseph said.
Instead of providing users with pages of links to search, AI Overviews streamlines the process by summarizing information and giving users a concise answer, Joseph says.
“They’ll experience less friction, they won’t have to click through to a bunch of different websites, which can often be quite a problem if you just want some information really quickly,” notes Chris Stokel-Walker, technology manager. journalist and author of the book “How AI Ate the World: A Brief History of Artificial Intelligence – and Its Long Future.”
MORE: Video — Google announces product updates at its developer conference
However, Stokel-Walker said the new feature makes it more difficult for people using Google Search to verify the accuracy of the information they read.
“Over the past 20 years, we’ve become accustomed to Google Search getting the dominant results from a search term that is largely correct,” he told ABC Audio. “If you get away from that, as Google is proposing, and basically just slide an answer straight into the search results page created via generative AI, suddenly you have no real way to identify and analyze that information to see if it’s true or not.”
There are also other concerns about the new feature. First, generative AI technology, both from Google and elsewhere, has been criticized for “hallucinating,” that is, generating information that is unreliable and inaccurate.
For example, in the few weeks since AI summaries have become available to the public, people using Google Search have been advised to eat at least one small stone a day, and one user was told that this is a good way to leave cheese on pizza sticking is to mix glue into the tomato sauce – both are obviously very bad ideas. It also said that Andrew Jackson, the seventh US president who died in 1845, graduated from college in 2005.
Stokel-Walker said any benefit from AI Overviews ultimately comes down to a trade-off between convenience and cost. “You no longer have to click through five or six different pages and perhaps multiple pages of search results to find the right answer, but it also means that the answer may be wrong or may not be the answer you are actually looking to get,” he said.
“We’ve always been very clear about the limitations of LLMs, that there will be occasional hallucinations,” Google’s Alex Joseph said. That’s why AI Overviews also mentions the websites it uses to generate its answers.
“It’s part of the reason we present all the information to you holistically,” says Joseph. “These are quick shortcuts that allow you to get some information quickly, but they are tracked along with areas where you can go, double check and verify.”
Joseph also said that not all searches are best served by an AI overview: “We only show them on searches that we are confident will be useful and actually improve the experience.”
MORE: Video Google announces product updates at its developer conference
In the wake of the unusual reactions reported by some social media users, Google announced that it has made “more than a dozen technical improvements” to AI Overviews. According to Liz Reid’s blog post, they include limiting the inclusion of user-generated content, as well as satirical or humorous web pages, in the data used to create AI summaries. Reid said Google has also “launched additional trigger refinements to improve our quality protections” regarding health content, and that it “aims[s] not to show AI overviews for hard news topics, where freshness and factuality are important.”
The blog post also notes that “AI overviews generally do not ‘hallucinate’ or make things up in the way that other LLM products might,” and that the incorrect answers are the result of “misinterpreting questions, misinterpreting a nuance of language on the internet, or because there isn’t a lot of good information available.”
Beyond accuracy concerns, Stokel-Walker said Google prioritizing AI summaries over traditional search results could impact revenue and reshape the way business is done on the internet.
“Websites produce content; they try to make it attractive to Google. Google will show them in the search results. And as a result, people click through to their website, then see ads on the back end of it, and the publisher makes the money they make can post new content on websites,” explains Stokel-Walker.
However, by replacing the top of Google’s search results page with AI-generated content, Stokel-Walker said websites could see fewer visitors – and therefore less advertising revenue.
According to Stokel-Walker, it is an ironic situation. That’s because Gemini, the LLM that Google uses to create its AI overviews, is dependent on the websites it now appears on.
“These websites still need to exist and they need to have a way to generate revenue, because otherwise there is nothing on which to base the AI-generated search results,” Stokel-Walker said.
MORE: Google joins forces to increase online transparency with a new digital standard
In a statement to ABC News, Google said its testing showed the opposite is happening: the links in AI Overviews are getting more clicks than if the page had appeared as usual in search results. Google also said it will “continue to focus” on sending valuable traffic to publishers and creators.
Regardless of how the concerns about AI overviews ultimately subside, it’s just one of many features the company has planned for its line of technology products.
“I think it’s concerning to do something as quickly as Google is doing it,” said technology journalist C. Scott Brown of the website Android Authority.
Google has announced plans for additional features similar to AI Overviews, which aim to answer questions about specific web pages or YouTube videos. Brown says these features will come to market against a backdrop of increasing competition.
“And the reason it does that is because it feels like it has to. It must keep pace with companies – especially like OpenAI, for example – that are creating generative AI technologies that threaten Google’s core business of delivering information to people. through Google Search, delivering them ads that could make Google billions and billions of dollars,” Brown said.
“Now that Google sees these things as a threat, it can’t just rest. It can’t figure out how to do this gracefully and properly,” Brown added. “It just has to do it.”
Google is adjusting AI overviews after a troubled rollout originally appeared on abcnews.go.com