In December 2025, a San Jose widow stopped a pig butchering scam in its tracks by asking ChatGPT one simple question. But you don’t need to wait until you’ve lost nearly a million dollars. Here’s how to use AI as your personal fraud detector—before you send a single dollar.

🎙️ Related Podcast: The Perimeter is Dead: How Vendor Insecurity Ignited a $500 Million Ransomware Crisis

The Moment That Changed Everything

Margaret Loke had already wired nearly $1 million to someone she’d never met. She’d drained her IRA, taken a second mortgage on her condo, and was being pressured to borrow more.

Then she did something that would save what little she had left: she described her situation to ChatGPT.

“ChatGPT told me: No, this is a scam, you’d better go to the police station,” Margaret told ABC7 News.

In seconds, the AI had identified patterns that months of emotional manipulation had obscured. The romantic approach, the cryptocurrency “investment,” the escalating requests, the frozen account requiring more deposits—these matched known fraud signatures that ChatGPT had been trained on.

For Margaret, it was a lifeline of clarity in an ocean of manipulation.

But here’s the thing: you don’t need to lose $1 million before using AI to check your situation. You can use it at the first suspicious message, the first investment pitch, the first “too good to be true” opportunity.

This guide will show you exactly how. ScamWatchHQ GPTIntroducing ScamWatchHQ: Your New AI-Powered Guardian Against Scams In today’s digital age, scams have become increasingly sophisticated, targeting individuals and businesses alike. From phishing emails and fraudulent phone calls to deceptive online advertisements, the variety of tactics used by scammers continues to expand. At ScamWatchHQ, we understand the emotional tollScamWatchHQScamWatchHQ

Why AI Works as a Scam Detector

Modern AI systems like ChatGPT, Claude, and Google Gemini have been trained on vast amounts of documentation about fraud patterns, scam playbooks, victim testimonials, and investigative journalism.

This means they can:

1. Pattern Recognition at Scale

AI can instantly compare your situation against thousands of documented scam cases. A human might know about a few fraud schemes; AI has processed millions of examples.

2. Emotional Neutrality

When you’re being scammed, you’re usually emotionally invested—whether in a romantic relationship, the excitement of potential profits, or the fear of consequences. AI doesn’t have emotional investment. It analyzes the facts.

3. No Judgment, No Embarrassment

Many scam victims don’t seek help because they’re embarrassed. They don’t want family members to know they sent money to someone online. AI doesn’t judge. It just analyzes.

4. Available 24/7

Scams often escalate outside business hours—late nights, weekends, holidays—when banks are closed and advisors are unavailable. AI is always there.

5. The Pause That Matters

The most important thing AI does might not be the analysis itself. It’s the pause it creates. When you type out your situation clearly to present it to an AI, you’re forced to step back, organize the facts, and see the situation from outside.

Sometimes that pause alone is enough to break through the manipulation.

The AI Scam Detection Toolkit: Which Tool to Use

ChatGPT (OpenAI)

Best for: General scam analysis, pattern matching, conversational follow-up questions

Access: Free tier available at chat.openai.com; ChatGPT Plus ($20/month) for GPT-4o

Strengths:

  • Widely available and well-known- Good at conversational analysis- Understands context well

Limitations:

  • Training data has cutoff dates (may not know about brand-new scam variants)- Can sometimes be overly cautious (“I can’t provide legal advice”)

Claude (Anthropic)

Best for: Long-form analysis, nuanced situations, detailed explanations

Access: Free tier available at claude.ai; Claude Pro ($20/month) for extended usage

Strengths:

  • Excellent at analyzing long conversations- Thoughtful, detailed responses- Good at explaining why something is suspicious

Limitations:

  • Smaller user base means fewer scam-specific training examples- May be more verbose than necessary

Google Gemini

Best for: Web search integration, checking current scam alerts

Access: Free tier available; Gemini Advanced ($20/month) for Gemini 1.5 Pro

Strengths:

  • Can search the web for current scam warnings- Good integration with Google ecosystem- Can check if a website or company has been reported

Limitations:

  • Younger platform with less scam-specific training- Sometimes provides less confident assessments

Our Recommendation

For most users: Start with ChatGPT or Claude (free tiers). Both are effective for scam detection.

For detailed analysis: Use Claude when you have long conversations to analyze or complex situations to explain.

For current information: Use Gemini when you want to check if a specific company, website, or phone number has been reported as fraudulent.

The Scam Detection Framework: What to Ask

Step 1: Describe the Situation Clearly

The quality of AI’s analysis depends on the information you provide. Include:

  • How you were contacted (dating app, social media, text, email, phone)- How long you’ve been in contact- What they’ve asked you to do (invest money, buy gift cards, send payments)- What reasons they’ve given (investment opportunity, emergency, legal trouble)- Any red flags you’ve noticed (excuses for not meeting, pressure to act quickly)- How much money is involved

Step 2: Use Clear, Specific Prompts

Generic prompts get generic answers. Here are effective prompts for common situations:

Romance Scam Detection

I met someone on [dating app/social media] [timeframe] ago. They claim to be 
[profession] living in [location]. We talk every day but they have refused 
to video chat, saying [excuse]. Recently, they've started talking about 
cryptocurrency investments and showed me screenshots of big profits. They 
want me to invest through a platform called [name]. They've asked me to 
keep our relationship and the investment private. Am I being scammed?

Investment Fraud Check

Someone is offering me an investment opportunity with [promised returns]. 
They want me to send money via [payment method] to [destination]. They say 
the investment is in [asset class]. They've shown me [evidence of returns]. 
I found them through [source]. They're pressuring me to invest before 
[deadline]. Is this legitimate?

Tech Support Scam Check

I received a [popup/call/email] claiming to be from [company]. They said my 
computer has [problem] and I need to [action]. They want me to [download 
software/call number/send payment]. They said if I don't act, [consequence]. 
Is this a scam?

Job Offer Scam Check

I received a job offer for [position] at [company]. The hiring process was 
[describe process]. They want me to [buy equipment/cash checks/provide 
personal info]. The pay is [amount] for [work described]. I found this job 
through [source]. Does this seem legitimate?

Step 3: Follow Up on Specifics

After the initial analysis, ask follow-up questions:

  • “What specific elements of my situation match known scam patterns?”- “What should I do next to verify this is legitimate or fraudulent?”- “What red flags did you identify that I should watch for in the future?”- “If this is a scam, what should I do now?”

Worked Examples: Real Scam Types, Real Analysis

Let’s walk through how AI analysis works for the most common scam categories.

Example 1: Pig Butchering (Romance + Crypto Investment)

What you might tell AI:

“I met a woman named Jessica on Facebook two months ago. She said she’s a financial analyst in Singapore. We message every day on WhatsApp—she sends good morning texts and asks about my day. She’s beautiful and seems to really care about me.

Last week she started talking about cryptocurrency trading. She showed me her trading account that has made over $200,000 in profits. She says she has a system that never loses. She wants me to try it with $5,000 on a platform called CryptoFuturePro. She says she’ll help me trade and we can build our future together.

I asked to video chat but she says her camera is broken. When I asked to meet she said she can’t travel right now because of work. Should I invest?”

What AI will identify:

✅ Classic pig butchering pattern: Romance building, followed by investment pivot ✅ Communication red flags: Daily texts, never video calls, always excuses for not meeting ✅ Investment red flags: Guaranteed returns, screenshots of profits (easily fabricated), unknown platform, mentor promising to “help you trade” ✅ Isolation tactics: Building exclusive relationship, implying shared future ✅ Likely trajectory: Initial “profits” will appear, followed by requests for larger investments, then account freeze requiring more deposits

What AI will recommend:

  • Do not invest any money- Reverse image search the photos (likely stolen from another person)- Search the platform name + “scam” (likely find reports)- Stop communication- If money was sent, report to FBI IC3

Example 2: Tech Support Scam

What you might tell AI:

“I got a popup on my computer saying Microsoft detected a virus and my banking information is at risk. There was a number to call. I called and a technician named Kevin said my computer is infected and hackers have my bank details.

He had me download a program called AnyDesk so he could look at my computer. He showed me a command prompt with red text that said ERROR and claimed this proves I’m hacked. He says I need to pay $399 for Microsoft Security Protection or the hackers will steal all my money. He wants me to pay with Best Buy gift cards. Should I pay?”

What AI will identify:

✅ Popup warning scam: Real antivirus software doesn’t display popups with phone numbers ✅ Remote access scam: AnyDesk is legitimate software, but scammers use it to take control ✅ Fake “evidence”: The command prompt “errors” are either normal system messages or commands the scammer typed ✅ Gift card payment: No legitimate company accepts payment in gift cards—this is pure money laundering ✅ Urgency/fear tactics: “Hackers will steal your money” pressures immediate action

What AI will recommend:

  • Do not pay anything- Hang up immediately- If AnyDesk was installed, uninstall it and run malware scans- Change passwords (from a different device if possible)- Microsoft never calls or pops up warnings asking you to call- Report to FTC at reportfraud.ftc.gov

Example 3: Job Scam

What you might tell AI:

“I applied for a remote customer service job and got hired after just an email interview. The HR person said they’ll send me a check for $4,500 to buy equipment from their vendor. I should deposit the check, then send $4,000 to the equipment company via Zelle, and keep $500 for my trouble.

The company is called TechSync Solutions and they have a website that looks professional. The salary is $45/hour for 25 hours a week. Is this real?”

What AI will identify:

✅ Fake check scam: Check will appear to clear, then bounce after you’ve sent the money ✅ Unrealistic hiring process: Legitimate jobs don’t hire after one email ✅ Payment forwarding: You’re being used as a money mule ✅ Too good compensation: $45/hour for entry-level remote customer service is far above market rate ✅ Website doesn’t prove legitimacy: Scam websites can look professional and be created in hours

What AI will recommend:

  • Do not deposit the check- Do not send any money- Research the company through independent sources (not links they provide)- Look up company on Better Business Bureau, Glassdoor, LinkedIn- Report to the job site where you found the listing- If check was deposited, contact your bank immediately

Example 4: Emergency/Grandparent Scam

What you might tell AI:

“I got a call from someone who sounded like my grandson crying. He said he was in a car accident in Mexico and is in jail. He said not to tell his parents because he’s embarrassed. A ‘lawyer’ got on the phone and said I need to send $8,000 bail money by Western Union today or he’ll be transferred to a dangerous prison. They gave me an address in Tijuana to send it.”

What AI will identify:

✅ Impersonation scam: Caller may have gotten family details from social media or the phone book ✅ Crisis scenario: Creates emotional panic that bypasses rational thinking ✅ Secrecy demand: “Don’t tell anyone” prevents verification ✅ Urgency: “Today or he’ll be transferred” prevents research ✅ Wire transfer: Untraceable payment method ✅ International element: Makes recovery impossible

What AI will recommend:

  • Hang up and call your grandson directly at his known number- Call his parents regardless of what the caller said- Real emergencies don’t require Western Union- Real lawyers don’t take bail payments themselves- Report to local police and FTC- Create a family code word for emergencies

Advanced Techniques: Getting Better Analysis

Technique 1: Paste Actual Messages

Instead of summarizing, paste the actual messages you’ve received. AI can analyze specific language patterns used by scammers:

Analyze these messages I received. Identify any patterns that suggest 
this might be a scam:

[Paste actual messages here]

Technique 2: Ask for the Specific Scam Type

If you suspect something is wrong but aren’t sure what:

Based on what I've described, what specific type of scam does this most 
closely resemble? What is the typical progression of this scam, and what 
will they likely ask me to do next?

Technique 3: Request Verification Steps

What specific steps can I take to verify whether this is legitimate? 
Give me concrete actions, not just "be careful."

Technique 4: Ask About Red Flags You Missed

Are there red flags in what I described that I might have missed or 
rationalized? What should have been warning signs earlier in this 
interaction?

Technique 5: Use AI to Help Report

I've determined this is a scam. What agencies should I report this to, 
and what information should I include in my reports?

What AI Can’t Do

1. Definitively Prove Legitimacy

AI can identify red flags and scam patterns, but it cannot definitively prove something is legitimate. A scam can have zero red flags initially; that doesn’t make it safe.

Use AI analysis as one input, not the only input.

2. Recover Lost Money

If you’ve already sent money, AI cannot help you get it back. It can help you identify appropriate reporting agencies, but recovery rates for scam losses remain extremely low.

Prevention is the only reliable strategy.

3. Replace Official Verification

AI can flag potential securities fraud, but it cannot verify whether a financial advisor is licensed. For that, you need:

  • FINRA BrokerCheck: brokercheck.finra.org- SEC EDGAR: sec.gov/edgar (verify company filings)- State insurance regulators: For insurance products

4. Know About Brand-New Scams

AI training has cutoff dates. A completely novel scam technique that emerged last week may not be in the training data. AI generalizes from known patterns, which usually works, but isn’t infallible.

5. Overcome Emotional Investment

If you’re truly convinced you’re in a legitimate relationship or investment, you might rationalize away AI’s warnings. The analysis only works if you’re genuinely open to the possibility you’re being scammed.

When to Consult AI: The Checklist

Use AI as a scam check whenever:

  • Someone you haven’t met in person asks for money- [ ] An investment promises guaranteed returns with no risk- [ ] You’re asked to pay via gift cards, wire transfer, or cryptocurrency- [ ] Someone creates urgency (“Act now or lose the opportunity”)- [ ] You’re asked to keep something secret from family/friends- [ ] A romantic interest pivots to discussing money or investments- [ ] You receive unexpected contact about a problem you didn’t know existed- [ ] Something feels “off” but you can’t articulate why- [ ] You’re about to send money to someone/somewhere you can’t verify- [ ] An offer seems significantly better than alternatives (too good to be true)

The cost of checking is zero. The cost of not checking can be everything you have.

Building Your Personal Anti-Scam Protocol

Step 1: The 24-Hour Rule

Before any financial action based on unsolicited contact, wait 24 hours. Use that time to:

  • Describe the situation to AI- Research the company/person independently- Consult with family or friends- Contact the organization directly through verified channels (not links/numbers you were given)

Step 2: The Second Opinion Protocol

Make it a rule: before sending money to any new recipient or making any unfamiliar investment, get a second opinion from:

  • AI analysis (free, instant, non-judgmental)- Trusted family member or friend- Financial advisor (for investment decisions)- Your bank (can sometimes flag suspicious transfers)

Step 3: The Verification Checklist

Before acting on any request for money:

  • Have I met this person in real life or verified video chat?- [ ] Can I find this company/person through independent research?- [ ] Does the payment method make sense? (Gift cards = always scam)- [ ] Why would a stranger be offering me this opportunity?- [ ] What happens if I wait 24 hours?- [ ] Have I asked AI to analyze this situation?

Step 4: The Code Word System

For family-based scams (grandparent scams, kidnapping hoaxes), establish a family code word that must be used in any emergency request for money. Share it only in person.

Helping Others: Sharing This Resource

Scam victims are disproportionately:

  • Seniors (especially those isolated or recently widowed)- People going through major life transitions (divorce, job loss, grief)- Those seeking companionship or financial opportunity- People unfamiliar with technology

If you have family members in these categories:

1. Introduce AI as a Tool, Not Surveillance

Don’t say: “I don’t trust your judgment.” Do say: “There are great AI tools now that can help spot scams. Want me to show you how they work?“

2. Make It Easy

  • Bookmark ChatGPT or Claude on their devices- Practice with them using hypothetical examples- Make asking AI the default, not a sign of suspicion

3. Create a Check-In Protocol

“Before you send money to anyone new, call me or check with ChatGPT—whichever is easier.”

4. Remove the Shame

Emphasize: “Millions of people get scammed. It’s not about intelligence—professional criminals spend all day perfecting these techniques. Smart people use tools to fight back.”

Resources for Scam Reporting and Recovery

Reporting Scams

FBI Internet Crime Complaint Center (IC3): ic3.gov For all internet-related fraud

FTC Report Fraud: reportfraud.ftc.gov For any consumer fraud

SEC Tips and Complaints: sec.gov/tcr For investment fraud

State Attorney General: Find yours at naag.org For local businesses and services

Social Security Administration OIG: oig.ssa.gov For Social Security-related scams

IRS Tax Scams: irs.gov/privacy-disclosure/report-phishing For IRS impersonation

Victim Support

Identity Theft Resource Center: idtheftcenter.org Free assistance and counseling

AARP Fraud Watch Network Helpline: 877-908-3360 Support for seniors

Financial Therapy Association: financialtherapyassociation.org Mental health support for financial trauma

Verification Resources

FINRA BrokerCheck: brokercheck.finra.org Verify financial advisors

SEC EDGAR: sec.gov/edgar Verify company filings

Better Business Bureau: bbb.org Business reputation checking

Scamadviser: scamadviser.com Website reputation checking

Conclusion: The New First Line of Defense

Margaret Loke’s story is dramatic: a last-minute save after nearly $1 million lost. But the real lesson isn’t about dramatic rescues—it’s about prevention.

Every scam victim was once at a moment where they could have stopped. Where they could have asked someone. Where they could have questioned what they were being told.

AI gives you that moment, on demand, without judgment, without embarrassment, without needing to explain to family members why you thought sending $50,000 to someone you’d never met might be a good idea.

It’s not a perfect tool. It can’t guarantee you’ll never be scammed. But it can:

  • Force you to articulate what’s happening- Identify patterns you might have missed- Provide emotional neutrality when you’re emotionally invested- Give you permission to doubt

That’s often enough.

The next time something seems too good to be true, the next time someone you’ve never met asks for money, the next time you feel pressure to act fast—

Stop. Open ChatGPT or Claude. Describe what’s happening. Ask: “Is this a scam?”

The answer might save everything you have.

This guide was prepared by the ScamWatch HQ research team. If this guide helped you avoid a scam or identify fraud, please share it with others who might benefit. Every person who checks with AI first is a person who doesn’t fund the criminal ecosystem with their life savings.