Google says it is manually removing wrong AI overview answers: Report

Google recently unveiled AI Overviews, a new feature for Search that offers Gemini-powered overviews of topics you are searching for. However, many users flagged issues saying they are seeing bizarre, false and unrelated AI-generated advice on Overviews like using non-toxic glue to stick cheese to pizza, plagiarizing text from blogs, and claiming that pythons are mammals.

Now, in an emailed statement to The Verge, Google spokesperson Meghann Farnsworth has confirmed that the tech giant is “taking swift action” to remove AI Overviews for some queries “and using these examples to develop broader improvements to our systems, some of which have already started to roll out.” Farnsworth also said that many of the examples generated by AI Overview were doctored and that they were unable to reproduce them.

Also Read | 6 reasons why OpenAI’s GPT-4o is the smartest AI assistant for everyday tasks

For those not in the loop, Google has been testing AI-generated overviews for almost a year now, with the feature made available for the first time in May last year as Search Generative Experience.

Story continues below this ad

Google CEO Sundar Pichai also claims that the company has already served more than a billion queries during this period, that the cost of showing AI answers has been reduced by 80 per cent, and that integrating AI into search will offer more value in the future.

In a recent short interview with the popular YouTuber Hayls World, Pichai opened up and explained why everyone should use Gemini and talked about the future of Google Assistant.