Google’s AI Overviews Can Scam You. Here’s How to Stay Safe

These days, rather Instead of showing you the traditional list of links when you run a search query, Google instead intends to release AI Previews: synthesized summaries of information pulled from the web, with a little word prediction magic added, and grouped in a way that appears as accurate and reliable as possible.

We’ve written before about some of the problems with these AI previews, which routinely contain errors or nonsense, and of course rip off the work of human editors who actually know the answers to the questions you ask Google. There’s another problem, though: these AI responses can actually be dangerous.

As with all new technologies in history, scams are now making their way into AI previews, apparently injecting Google’s AI responses with fraudulent phone numbers that you shouldn’t trust. Here’s what’s happening and how you can stay safe.

How AI Showcase Scams Work

It’s a good idea not to trust AI with coordinates.David Nield

The Washington Post and Digital Trends spotted instances of fraudulent support numbers appearing in Google’s AI previews, reports of which appeared on Facebook and Reddit, respectively. Credit unions and banks also warn their customers about these scams.

This doesn’t seem like a completely new problem, but the way Google Search works today has taken a new turn.

Here’s what happens: the unfortunate victim Googles the name of a company looking for a contact number, then calls the number displayed by the AI. This doesn’t actually lead to the company in question, but rather to someone pretending to be that company, who then attempts to obtain payment information or other sensitive details from the caller.

It’s unclear exactly how these fake numbers are introduced, but the best guess is that they are published on several discreet online sites, alongside the names of major companies. AI Overviews then comes in and grabs them, without doing the proper checks to verify the information.

The use of deceptive phone numbers by bad actors is of course not a completely new danger; disinformation has been part of the web for a long, long time. But the design of AI previews, which cherry-pick information from the web and present it as fact rather than encouraging you to do the research yourself, makes people much more vulnerable to this type of scam.