Only a few days into the debut of AI Overviews in Google Search, and things are already not going well for the company. Introduced in the US at the moment, queries are throwing up inaccurate results within the AI feature.

AI Overviews show a quick set of answers to search questions at the very top of Google Search. For example, if you search for the best way to bake a chocolate cake, the results page may show an AI Overview at the top with a step by step process, combing information it synthesized from around the web.

How bad have been the results? To a query about daily “rock intake” for a person, AI Overview suggested we eat “at least one small rock per day”. That advice hailed from a 2021 story in The Onion.

It’s a variation on AI hallucinations that occur when a generative AI model serves up false or misleading information and presents it as fact.

It’s not the first time that Google is having AI-related problems. In February 2023, when Google announced Bard, a chatbot to take on ChatGPT, it shared incorrect information about outer space. And this February, the company released Bard’s successor, Gemini, a chatbot that could generate images, users found that the system refused to generate images of white people in most instances and offered inaccurate depictions of historical figures.

“The vast majority of AI Overviews provide high quality information, with links to dig deeper on the web,” a Google spokesperson told CNBC in a statement. “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.” The company is taking “swift action where appropriate under our content policies”.

According to Cnet, the search engine that debuted in 1998 controls about 86 per cent of the market. Google’s competitors don’t come close: Bing controls 8.2 per cent, Yahoo has 2.6 per cent, DuckDuckGo is 2.1 per cent, Yandex has 0.2 per cent and AOL is 0.1 per cent.