A restaurant owner in Montana is pleading with customers to stop using Google's artificial intelligence (AI) overview to inquire about daily specials, as the system keeps providing incorrect information, leading to angry customers yelling at staff.
According to First Alert4, Eva Gannon, the owner of Stefanina's Wentzville in Montana, said that Google's AI system often pushes false offers to customers, even fabricating entire menu items. For example, the system once incorrectly told customers that the restaurant sold large pizzas at the price of small ones.
"As a small business, we can't fulfill the special offers from Google AI," Gannon told the local television station helplessly. She also added, "We can't control the content or statements released by Google." The restaurant clearly stated in a strong Facebook post that Google AI was inaccurate and urged customers to visit their Facebook page or website for correct information.
This phenomenon is called "hallucination," a common flaw in AI chatbots and large language models, where they generate seemingly reasonable but completely wrong information. Although Google's AI overview has been mocked for its instability (such as suggesting people put glue on pizza), many users still seem unaware of the inherent unreliability of these tools.
Stefanina's Wentzville is not the only victim. In June this year, a solar energy company in Minnesota sued Google for defamation, claiming that its AI overview system spread false information about the company, damaging its business. The system once incorrectly claimed that the company was facing a lawsuit over fraudulent sales practices.
Although Google is actively promoting its AI mode, encouraging users to "ask Google first," and announcing that the AI mode can now help users book restaurants, the above cases show that users should not fully rely on these AI tools to obtain factual information or make decisions.
