Follow Slashdot blog updates by subscribing to our blog RSS feed

 


Forgot your password?
Close
typodupeerror
AI Crime Google

Google's 'AI Overview' Pointed Him to a Customer Service Number. It Was a Scam (yahoo.com) 54

Posted by EditorDavid from the thanks-for-nothing dept.
A real estate developer searched Google for a cruise ship company's customer service number, reports the Washington Post, calling the number in Google's AI Overview. "He chatted with a knowledgeable representative and provided his credit card details," the Post's reporter notes — but the next day he "saw fishy credit card charges and realized that he'd been fooled by an impostor for Royal Caribbean customer service."

And the Post's reporter found the same phone number "appearing to impersonate other cruise company hotlines and popping up in Google and ChatGPT" (including Disney and Carnival's Princess line): He'd encountered an apparent AI twist on a classic scam targeting travelers and others searching Google for customer help lines of airlines and other businesses... The rep knew the cost and pickup locations for Royal Caribbean shuttles in Venice. [And "had persuasive explanations" when questioned about paying certain fees and gratuities.] The rep offered to waive the shuttle fees...

Here's how a scam like this typically works: Bad guys write on online review sites, message boards and other websites claiming that a number they control belongs to a company's customer service center. When you search Google, its technology looks for clues to relevant and credible information, including online advice. If scammer-controlled numbers are repeated as truth often enough online, Google may suggest them to people searching for a business.

Google is a patsy for scammers — and we're the ultimate victims. Google's AI Overviews and OpenAI's ChatGPT may use similar clues as Google's search engine to spit out information gleaned from the web. That makes them new AI patsies for the old impostor number scams.
"I've seen so many versions of similar trickery targeting Google users that I largely blame the company for not doing enough to safeguard its essential gateway to information," the reporter concludes, (adding "So did two experts in Google's inner workings.") The Post is now advising its reader to "be suspicious of phone numbers in Google results or in chatbots."

Reached for comment, a Google spokesman told the Post they'd "taken action" on several impostor numbers identified by the reporter. That spokesman also said Google continues to "work on broader improvements" to "address rarer queries like these." OpenAI said that many of the webpages that ChatGPT referenced with the bogus cruise number appear to have been removed, and that it can take time for its information to update "after abusive content is removed at the source."
Meanwhile, the man with the bogus charges has now canceled his credit card, the Post reports, with the charges being reversed. Reflecting on his experience, he tells the Post's readers "I can't believe that I fell for it. Be careful."

Google's 'AI Overview' Pointed Him to a Customer Service Number. It Was a Scam More | Reply

Google's 'AI Overview' Pointed Him to a Customer Service Number. It Was a Scam

Comments Filter:

  • ok? (Score:1)

    by Anonymous Coward
    Did he miss the news stories about how the very same AI service was telling people to eat rocks and to put glue on pizza? He doesn't find it offensively wrong 90% of the time?

    • Re:ok? (Score:5, Insightful)

      by Z00L00K ( 682162 ) on Sunday August 17, 2025 @11:39PM (#65596526) Homepage Journal

      The AI is more wrong than right in many searches that should be simple so I have started to totally ignore those results and not even do a thumbs down because it's not worth my time to teach that AI what it does right or wrong.

      • https://github.com/zbarnz/Goog... [github.com]

        You're even more right than you know about the thumbs down because the AI doesn't learn anyway.

    • Is my experience that AI is only wrong 10% of the time (e.g. when it regurgitates mainstream economic theories) unique?

      Why shouldn't the scammed check the number first? Why does that reporter excuse laziness in not doing the slightest bit of due diligence?

      • Check the number where? I'm Google? With ChatGPT? The same ones that told them the number?

        We use Google to check the phone numbers!

        • Re: (Score:1, Troll)

          by bleedingobvious ( 6265230 )

          Does the company in question have a verified public portal?

          Why, yes. Yes it does. Why use anything other than that to contact the company?!

          Oh. That's right. Idiots are no longer responsible for their careless idiocy anymore. Fark this dbag. Well earned own goal.

          • Re: (Score:2)

            by moranar ( 632206 )

            So where and how should people find this verified public portal?

            If you're about to say "you can use the exact same method they used to find the phone number" and you don't see the issue, you have a problem.

            • If you're about to say "you can use the exact same method they used to find the phone number" and you don't see the issue, you have a problem.

              i) There are more search engine options

              ii) If you can't be bothered to check the domain and cert, you're just another cry-baby waiting to bleat publicly about how hard you got pwned

              • Re: (Score:2)

                by moranar ( 632206 )

                Oh, so now any random person should know and use and contrast multiple search engine results for routine searches, good, good, that will definitely happen. And common people definitely do check website security certs, which are also not a thing that anybody can get for their random domain. And anybody who falls for it is a crybaby. Great insight.

                I don't know where you got that worldview, but I wish you could reevaluate.

    • Re:ok? (Score:5, Funny)

      by ebunga ( 95613 ) on Monday August 18, 2025 @12:57AM (#65596594)

      But "everyone" says it's "getting better" and doing it "very very fast" and that I am very smart and that the AI is a god. All hail the AI.

      • Re: (Score:3)

        by TWX ( 665546 )

        I'm reminded of a joke meme...

        Interviewer starts, "so it states here on your resume that you can do multiplication very, very fast. So what's 131x467?"

        The applicant replies, "76,543!"

        The interviewer objects, "that isn't even close to right!"

        The applicant responds, "well, it was very, very fast!"

    • Re: (Score:3)

      by Junta ( 36770 )

      Every story is accompanied by a bunch of white knighting saying that it's fake, the user was deliberately trying to induce funny responses, or that it's fixed now.

      If your typical google search is pretty milquetoast, then it's not crazy to imagine that the results are inoffensively 'ok', and that maybe he tried to ask about eating rocks and pizza and got the post-correction behavior and thus concluded the people dismissing the criticism were right.

      He might have felt that it may not get things totally right,

    • Google keeps putting AI summary before the search results. They are operating a deceptive business just to push AI slop onto people who don't want it.

    • Did you miss the relentless propaganda from the AI industry, posted here and in every news publication, about how AI is so awesome it'll replace most jobs and you can do anything with it? Did you miss the inherent propaganda of Google putting Gemini results on nearly every search page?

  • I travel a fair bit for work, and if I'm going to a new (to me) city I'll often google "X hotel chain Y city" to see what options I have. Unfortunately Google has been victimized in their search results on this for a long time; it's not just redirects to various third-party hotel sites but it can be other sites entirely as well. This can lead to outdated or completely inaccurate information, and bad phone numbers too. Sometimes I'll see the correct link about 5-6 entries down, but that's not guaranteed - and if chain X has more than one hotel of name A in or near city Y the results could well be missing some locations on the first page.

    This is the old googlebombing technique from years ago (remember "miserable failure"?), but deployed for profit instead of politics.

  • Well duh (Score:5, Insightful)

    by locater16 ( 2326718 ) on Sunday August 17, 2025 @11:48PM (#65596542)
    Google was already SEO'd to absolute shit for search, why shouldn't the same happen in record time for AI?

    • Re: (Score:1)

      by Anonymous Coward

      The problem is that they're trying to mimic a fundamentally flawed design. The whole point of AI should be that it's better than humans, not copying them poorly.

      Current "AI" can't even follow basic logic. A computer that can't follow logic. WTF is the point of that?! We already have humans for that.

    • It's kind of the opposite process for SEOing AI. Instead of creating a site that Google sees is better than everything else, Google tells you what's the best url/phone and then you go register that.

      But this isn't new. Phone numbers from online search, review sites, and index sites were already suspect. There's tons of old stories of how scammers (including gig companies) find companies which haven't registered on those sites then registers themselves and their contact info, pretending to be that company.

  • AI is getting better, but (Score:4, Interesting)

    by MpVpRb ( 1423381 ) on Monday August 18, 2025 @12:47AM (#65596582)

    Always cross check everything
    This is why I like Perplexity. It gives me links to sources

    • Re: (Score:1)

      by Anonymous Coward
      Do you personally check the sources?

  • I remember talking with my mom about some trip she was taking and she mentioned going online to have USPS hold her mail while she was gone, and having to pay for it. It wasn't much, something like $15, but she had her mail held as requested and it didn't appear like anyone charged anything but the specified fee.

    The USPS doesn't charge anything for requests to hold mail, forward mail, or similar kinds of adjustments to mail delivery. At least I didn't find any indication they charged fees for that. Plenty

  • Could Google use AI for fraud detection? (Score:5, Insightful)

    by misnohmer ( 1636461 ) on Monday August 18, 2025 @02:20AM (#65596660)
    AI should be able to "investigate" any phone number it's going to be publishing in Google results. Check who is the owner of said number, does it match the company's website, heck, have AI call the number to attempt to verify it (perhaps ask the operator who picks up to provide a callback number/extension which would match up with any of the official company numbers). Sure, some companies with shitty IT would get flagged as fake, but if Google flags their number as fake, they'll be fixing it in no time - no different than lowest bidder IT setting mail servers which end up getting flagged as spam servers (sometimes because they are easy to exploit and in fact become spam servers), once a company realizes all of its emails go into junk folders of major email providers (Google, Microsoft), they prioritize fixing their email servers (or often just outsource it to Google or Microsoft).
    • thers a reason it says ai can be wrong make sure you check.
    • If Google's AI tried to call every number it encounters online, I think the number making all the unsolicited calls would quickly be blacklisted and reported to do-not-call lists and sites.

    • Check who is the owner of said number, does it match the company's website, heck, have AI call the number to attempt to verify it (perhaps ask the operator who picks up to provide a callback number/extension which would match up with any of the official company numbers).

      So every time Google generates an AI summary that includes a phone number—how many times a second do you reckon that happens?—you want an automated robocall to spam whatever business or random person’s phone number happen

      • No, not every time it generates a summary, but rather every time it encounters a new number enough times that it might be weighted heavily enough to show up in an AI summary. Once validated, no need to re-validate it. For example 1-800-GOT-JUNK would be validated once, and then included in millions of AI summaries per second if needed, without having to re-validate it.

  • their Googly AI system for a list of scammer websites. Can then block them.

    Eeezee-Peezee

  • This happened to me with delta airlines. I thought through the conversation I had and called the number back and got the exact same rep and I knew it was a scam and cancelled the card immediately. Good thing is the transaction hadn't gone through and I didn't loose anything. Don't trust google links.

  • The real problem is companies making it impossible to contact them about anything. Give me an easy to get to phone number and chat box option. Even a I want to send a message option seams to be disappearing, just an endless loop of pages that don't help that then point you to contact them that is really just a feed back into the loop of unhelpful help articles.

    If the company had presented an easy phone number, the AI or search results would have given it.

  • Phone numbers are cheap but if the bad guy isn't spending money on identifying what the victim is thinking, it means the victim at some point has to say "Is your business name, XYZ?", which a potential customer should not be doing.

    Presumably, this guy wasn't comparison shopping. That tends to indicate what fees (carriage/transfer/holding costs, insurance, city/state taxes, deposit) are normal for a category of service.

Slashdot Top Deals

Like punning, programming is a play on words.

Close