ai-security

Why Are AI Chatbots Leaking Your Phone Number?

AI chatbots are handing out real people's private phone numbers to total strangers — not through hacking, but through a flaw in how they were trained. One Reddit user started getting calls from people looking for a lawyer, a locksmith, and a product designer — all because Google's AI decided their n

Lucas Oriens Kim

21 5월 2026 • 9 min read

Quick Answer
AI chatbots — including Google's — are accidentally surfacing real people's private phone numbers when users ask them to find contact info for businesses or services. Your number could already be circulating to strangers right now, and you'd have no way of knowing until your phone starts ringing with calls meant for someone else — or worse, someone who wants to scam you.

The Real Case: A Reddit User's Phone Turned Into a Switchboard

Millions of phone numbers scraped from public web sources are embedded in major AI training datasets, with no opt-out mechanism currently available to individuals.

This isn't theoretical. A Reddit user recently went public with a nightmare that started with a flood of bizarre phone calls. Strangers were ringing them looking for a lawyer. Then a locksmith. Then a product designer. The calls kept coming — all from people who had asked Google's AI chatbot for a phone number and received theirs instead.

Google's AI had apparently absorbed the victim's phone number from somewhere on the web and began serving it up as a contact number for unrelated businesses. The user had no relationship to any of these services. They just had the misfortune of having their number exist somewhere in the data the AI was trained on.

This is what researchers and privacy advocates are now calling 'AI-assisted doxxing' — except the AI isn't being malicious. It's just confidently wrong in a way that exposes real people to a firehose of strangers. And here's what makes this genuinely dangerous: fraudsters are already paying attention. A scammer who knows your number, your name, and that you're being mistaken for a service provider has everything they need to craft a targeted attack.

💡 Key Insight: The AI isn't hacking you — it's just treating your phone number like a public fact, which makes it almost impossible to fight.

How AI Chatbots End Up Leaking Your Number

Google's AI Overviews now appear in over 25% of all search results, meaning AI-generated contact information reaches hundreds of millions of queries daily.

The mechanics here are less glamorous than a cyberattack, and that's exactly what makes them so hard to stop. Here's what's actually happening:

1. **Your number gets posted publicly.** Maybe you listed it on a business directory, a forum, a classified ad, a church website, or a LinkedIn profile years ago. You've long since forgotten it.

2. **An AI training crawler indexes it.** Companies like Google, OpenAI, and others train their models on massive sweeps of the public internet. Your number gets absorbed alongside the surrounding text — maybe near keywords like 'plumber,' 'attorney,' or 'repair.'

3. **The AI makes an association.** The model doesn't 'know' your number is private. It learned that numbers appear near certain types of businesses, so it treats yours as a legitimate answer when someone asks for a contact.

4. **Someone asks the chatbot for a phone number.** 'What's the number for a locksmith in [your city]?' The AI responds with yours — confidently, fluently, incorrectly.

5. **Strangers call you.** Most are just confused. But some are fraudsters who now have a verified live number attached to a real person.

The cruelest part: the AI has no mechanism to check whether the number is still accurate, private, or belongs to someone else. It just answers.

💡 Key Insight: One old forum post with your number on it could be feeding stranger's phone calls to you right now — years after you posted it.

Why This Is More Dangerous Than It Sounds

The FTC received over 2.6 million fraud reports in 2023, with phone-based scams accounting for the highest per-incident dollar losses of any contact method.

Most people hear this story and think: 'Okay, annoying. Wrong number calls. Whatever.' That's the wrong reaction.

Here's the counterintuitive risk that guides aren't talking about: the people most likely to misuse AI-surfaced phone numbers aren't random scammers cold-dialing lists. They're targeted fraudsters who use the AI to build a profile on you first.

Think about it. If a scammer asks an AI chatbot for your number and gets it, they may also be asking: 'What does [your name] do professionally?' 'Is [your name] associated with any businesses?' 'What neighborhood is [business name] in?' AI chatbots will answer all of those questions using public data — and suddenly the scammer has your number, your approximate location, your professional background, and a pretext for calling you.

This is called 'social engineering enrichment,' and it's what turns a wrong-number glitch into a fraud pipeline.

There's also a psychological trap here: if someone calls you claiming to be a business you're apparently associated with in the AI's output, you'll be confused and disoriented. Confusion lowers your defenses. Scammers know this — disorientation is a tool, not an accident.

💡 Key Insight: Confusion is a weapon — and being mistaken for a business makes you easier to social-engineer, not just annoying to deal with.

Your Defense Checklist: What to Do Right Now

Data broker opt-out requests can take 30–90 days to process, meaning the window to act is now — not after the calls start.

Stop waiting for Google to fix this. Here's what you can actually do today:

**1. Google your own phone number immediately.** Put your number in quotes: `"555-867-5309"`. See every site that's published it. This takes three minutes and most people have never done it.

**2. Submit removal requests to data brokers.** Sites like Spokeo, WhitePages, BeenVerified, and Intelius are major sources for AI training data. Use their opt-out pages directly — or pay for a service like DeleteMe ($129/year) or Incogni (~$97/year) to do it automatically across 100+ brokers.

**3. Request removal from Google's search index.** Google has a 'Results About You' tool (myaccount.google.com/data-and-privacy) that lets you flag personal contact info for removal from search. Use it. It won't scrub the AI's memory instantly, but it reduces future exposure.

**4. Set up a Google Voice number for any public-facing use.** Give out the Google Voice number for classifieds, forums, or business directories. Keep your real number private. If the Voice number gets flooded, you can replace it without touching your actual line.

**5. If strange calls start, don't engage — document.** Note the date, time, and what the caller was looking for. Report to the FTC at ReportFraud.ftc.gov. If the pattern looks targeted rather than random, contact your carrier about temporary call filtering.

If you're still using your real number on public directories and haven't checked your Google search footprint, you're not protected — you're just lucky so far.

💡 Key Insight: A $129/year data broker removal service is cheaper than one afternoon dealing with fraud fallout.

Key Takeaways

🎯A Reddit user's phone number was served by Google's AI as a contact for at least three unrelated businesses — lawyer, locksmith, product designer — triggering a flood of stranger calls with zero warning.

📌AI chatbots don't distinguish between 'public data' and 'private data' — if your number appeared anywhere online, it may already be embedded in model outputs being served to millions of users daily.

⚡The real threat isn't wrong-number calls — it's that fraudsters are using AI chatbots to build detailed profiles on targets before calling, turning a data glitch into a targeted scam setup.

🔑Google your phone number in quotes right now, then submit opt-out requests to Spokeo, WhitePages, and BeenVerified — this is the single highest-leverage action you can take in under 20 minutes.

💎This problem will get worse before it gets better: AI models are retrained infrequently, meaning even if your data is removed from the web today, it may stay embedded in existing model weights for 12–24 months.

FAQ

Q: How do I find out if an AI chatbot is already sharing my phone number?
A: Search your phone number in quotes on Google, Bing, and also paste it directly into ChatGPT or Google Gemini asking 'what business is associated with this number?' — the results will tell you fast. If any chatbot returns a business name you're not connected to, you have confirmation and should start removal requests immediately.

Q: Can I force Google or OpenAI to remove my number from their AI's responses?
A: Honestly, this is the hard part — there's no direct 'delete me from the AI' button yet, and model retraining cycles mean your data can persist for months even after source pages are removed. Your best lever is Google's 'Results About You' tool for search removal, combined with data broker opt-outs to reduce the underlying sources the AI originally learned from.

Q: What should I do if I'm already getting strange calls meant for another business?
A: Document every call — date, time, what the caller wanted — and report the pattern to the FTC at ReportFraud.ftc.gov, then file a complaint directly with Google via their Search Console removal tool. If the calls feel targeted rather than random, tell your carrier and enable their built-in spam filtering (T-Mobile Scam Shield, AT&T ActiveArmor, and Verizon Call Filter are all free at the basic tier).

Conclusion

Your phone number is not as private as you think, and AI chatbots are now one of the fastest ways it ends up in a stranger's hands. The fix isn't complicated — but it requires you to act before the calls start, not after. Search your number in quotes on Google right now, then spend 15 minutes submitting opt-out requests to the three biggest data brokers: Spokeo, WhitePages, and BeenVerified. That one session could be the difference between a normal Tuesday and a week of fielding calls from people who think you're their locksmith.

How Does AI Help Cybersecurity Teams — And How Do Attackers Abuse the Same Tools?
The same AI tools that help security teams detect threats in milliseconds are being weaponized to clone voices, generate perfect phishing emails, and impersonate executives on live video calls. This isn't a future risk — it already cost one company $25 million in a single afternoon. Here's exactly w
How Did 35,000 Users Get Hit by Phishing?
Between April 14–16, 2026, attackers posing as HR departments stole authentication tokens from 35,000 people across 26 countries — without ever needing their passwords. Standard two-factor authentication didn't stop it. Here's what actually will.
How Are AI Deepfakes Used in Romance Scams?
Romance scammers are now using real-time AI deepfake video and cloned voices to impersonate attractive strangers — and sometimes even your own family members. The technology costs less than $20/month and is shockingly convincing. Here's what the attack looks like and how to protect yourself today.