‘It’s coming again on us’

WENTZVILLE, Mo. (KMOV/Gray News) – A restaurant in Missouri mentioned a synthetic intelligence instrument designed to reinforce Google search outcomes is inflicting confusion for patrons.

Stefanina’s Pizzeria in Wentzville posted to social media to warn patrons about search outcomes which will embrace deceptive or outright false details about their specials.

“Please do not use Google AI to find out our specials,” the publish learn. “Please go to our Facebook page or our website. Google AI is not accurate and is telling people specials that do not exist, which is causing angry customers to yell at our employees. We cannot control what Google posts or says, and we will not honor the Google AI specials.”

Eva Gannon is a member of the household that owns the restaurant. She mentioned the problem they’ve been dealing with is that clients informed them they’d discovered about offers or gives from Google that the restaurant was not actually providing.

One instance she shared was an AI Overview which confirmed that Stefanina’s would supply a big pizza for a similar value as a small one.

She additionally mentioned sure menu gadgets generated by the instrument have been incorrect.

“It’s coming back on us,” Gannon mentioned. “As a small business, we can’t honor a Google AI special.”

KMOV reached out to Google to study extra concerning the state of affairs however mentioned they haven’t but heard again.

However, the corporate’s information to utilizing AI warns customers that outcomes might typically be inaccurate because of the instrument making errors or misunderstanding the information it’s looking out by way of.

Jonathan Hanahan is a Washington University Professor who makes use of and teaches synthetic intelligence. He described it as a instrument that, whereas helpful, nonetheless requires skepticism by customers and a willingness for them to fact-check it.

“There’s a big difference between using AI and working with AI,” he mentioned.

Hanahan used a golf scoring app he has been engaged on for instance. Using AI, he can shortly incorporate course data which may usually take him hours to assemble. However, for the app to be efficient, he mentioned he nonetheless must cross-check the knowledge the software program is gathering.

Hanahan mentioned that a difficulty with utilizing AI to seek for data is that sure wording within the immediate might skew data, risking affirmation bias.

“It will sometimes take liberties to get you what you’re looking for,” he mentioned.

Hanahan mentioned the takeaway for customers must be to consider carefully concerning the sorts of questions and prompts they’re utilizing to assemble data and to keep away from what he described as a binary relationship.

“Think about how you’re asking prompts? How are you asking questions?” he mentioned. “Thinking about the greater conversation and not just the things you’re expecting to find out.”

Copyright 2025 KMOV by way of Gray Local Media, Inc. All rights reserved.