Episode Transcript
[00:00:00] Speaker A: Imagine getting a message from someone who remembers your dog's name, asks about your mother's surgery, checks in after a bad day at work, and somehow always knows exactly what to say.
Not too much, not too little, just enough to feel human.
[00:00:17] Speaker B: And now imagine that person is not a person at all. It is a chatbot patient, available, flattering, and designed to make you trust it before it ever asks for a dime.
[00:00:30] Speaker A: That is the new frontier of social engineering.
The old scam email with bad grammar and a suspicious prince is not gone, but it has been joined by something more dangerous. Artificial intelligence that can hold thousands of conversations at once and make every one of them feel personal.
[00:00:47] Speaker B: Thousands of fake friendships, fake romances, fake mentorships, fake business partnerships. And each one can be nurtured around the clock. No coffee breaks, no typos unless the bot is told to add them. No forgetting your birthday, which is unsettling
[00:01:06] Speaker A: because remembering birthdays used to be how you proved you cared.
Now it may just prove someone has a spreadsheet and a very obedient language model.
[00:01:15] Speaker B: To be fair, Nick, a spreadsheet remembering birthdays would still outperform you in certain family situations.
[00:01:21] Speaker A: I object to that characterization.
I remember the important ones. Yours is strategically placed right before Christmas. And mine, by the way, is right after Christmas. And Sue, I remember when we first met, you said I was too good to be true and that I had to be some kind of chatbot listeners.
[00:01:41] Speaker B: If this episode suddenly ends, you will know why.
[00:01:45] Speaker A: Okay, sue, back to the show.
But in my years working financial crimes, I learned that fraud is rarely about the money at first. It is about access. It is about lowering the victim's guard. The money comes later, after trust has been manufactured.
[00:02:01] Speaker B: Today on behind the Scams, we are looking at AI social engineering, the long con, where a machine builds the relationship and a criminal steps in only when it is time to cash out.
[00:02:14] Speaker A: And the numbers tell us this is not some niche corner of cybercrime. The FBI describes cryptocurrency, investment fraud, widely known as pig butchering, as one of today's most prevalent and damaging fraud schemes. Its operation, Level up, was created to identify victims early, and the Bureau said it had notified more than 8,000 victims, many of whom reportedly did not even know they were being scammed when agents reached them.
[00:02:40] Speaker B: That detail stops me cold, because it means the relationship can feel so plausible, so normal, that even an outside law enforcement call sounds like the unbelievable part.
The scammer has become the trusted voice.
The warning becomes the interruption.
[00:02:58] Speaker A: And before anyone says, well, I would never fall For a chatbot, Let me put on my old badge voice for a second.
Confidence is not a fraud control strategy in financial crimes. The people who worried me most were not the cautious ones. They were the ones who were certain they could spot a scam on instinct. Scammers study instinct. AI lets them test, refine and personalize the approach faster than any human crew ever could.
[00:03:25] Speaker B: So this episode is not about panic. It is about recalibrating trust.
Because the Internet has trained us to accept relationships that begin with text, AI has now trained criminals to make that text feel warm, specific and responsive. And once a machine can simulate intimacy, the old advice look for bad grammar or watch for weird phrasing starts to feel like bringing a flashlight to a house fire.
[00:03:54] Speaker A: We are going to walk through five how the relationship starts, how the money enters, how the technology supplies fake proof, why the aftermath is so psychologically punishing, and what listeners can do before or after the trap springs. We are also going to look at real cases where AI generated faces. Chatbot style messaging, voice cloning, synthetic video or AI assisted translation changed the victim's sense of reality.
Because that is the heart of this. Scammers are not just stealing money, they are they are attacking the evidence people use to decide what is real.
[00:04:31] Speaker B: To make this concrete, we are going to follow a fictional composite victim we will call Ellen. She is not one real person and her story is not meant to stand in for every victim.
She is built from patterns. Investigators, banks, families and victims describe again and again a competent adult, recently stretched thin, digitally connected, and just human enough to answer a kind message on a hard day.
[00:05:02] Speaker A: Ellen is 61. She works in administration for a regional medical group, knows her way around spreadsheets, pays her bills on time, and has raised two kids who now send her articles about scams which she pretends to appreciate and then mostly ignores.
She is not reckless. She is tired.
Her husband died three years ago. Her daughter lives two states away.
Most evenings Ellen eats dinner standing at the kitchen counter while scrolling through her phone.
[00:05:30] Speaker B: One night, after she comments on a post about retirement planning, she gets a message from a man named Daniel.
His profile says he is a widower, a civil engineer originally from Oregon, currently consulting overseas.
He does not flirt. At first he asks a normal question about the post.
Then he asks how long she has worked in healthcare.
Then he compliments her practical advice. It is ordinary enough to feel safe.
[00:06:01] Speaker A: That is the point of the opening. Daniel is not designed to sweep Ellen off her feet. He is designed to enter her routine. A chatbot can handle that perfectly. Modest questions, quick replies, gentle humor, no pressure. If Ellen answers at 6:40 in the morning before work, Daniel answers at 6:42. If she writes at 10:15 at night because the house feels too quiet, Daniel is awake then too.
[00:06:27] Speaker B: And let's be honest, if someone replies at 6:42 in the morning with warmth and correct punctuation, that already puts them in the top 2% of digital communicators.
[00:06:38] Speaker A: True, but that is exactly why it works. We are so used to rushed texts, thumbs up replies and autocorrect disasters that a calm, attentive stranger can feel like a minor miracle.
[00:06:50] Speaker B: Or like someone with no job, no hobbies and suspiciously excellent bandwidth. Let's start with where this begins. A dating app, a LinkedIn message, a wrong number text, WhatsApp telegram somewhere people are used to talking with strangers but not expecting a criminal operation behind the curtain.
[00:07:13] Speaker A: The first message is usually harmless. Sorry, wrong number. I liked your post.
We met at the conference, didn't we?
The goal is not to steal money on day one. The goal is to get a reply. In law enforcement we used to call that opening the door.
Once the door is open, the scammer starts mapping the room.
[00:07:33] Speaker B: And now the mapping can be automated. Large language models can keep a conversation going, mirror someone's tone, ask thoughtful follow up questions, and remember details. If you mention your golden retriever, your divorce, your new job, or your fear that your kids think you are slipping, that information becomes emotional.
[00:07:54] Speaker A: Inventory security researchers have documented how these operations use AI to generate short replies, translate back and forth between languages and tailor tone. Flirty, professional, sympathetic, urgent, depending on the target.
OpenAI, for example, reported banning accounts that appeared to originate in Cambodia and were being used to translate and generate romance baiting and investment scam messages in Japanese, Chinese and English.
That is not science fiction. That is a criminal call center. Discovering productivity Software and what did those
[00:08:30] Speaker B: operators allegedly do with the AI generated language?
They commented on real people's social posts, sometimes old posts, often targeting men whose profiles suggested they were over 40, including medical professionals. They were not blasting generic spam. They were fishing in specific ponds with bait that looked casual. A compliment here, a question there. A little human seeming curiosity.
[00:08:59] Speaker A: The targeting matters.
A doctor or engineer may be less likely to click a crude email link, but they may respond to someone who seems educated, worldly and interested in the same hobbies. Golf, travel, finance. A humanitarian cause. The first move is not the scam. The first move is compatibility. That is the part that makes this feel different.
The bot does not need to be brilliant. It needs to be Consistent. It needs to ask, how did the doctor's appointment go three days later? It needs to say, I know this week has been hard.
People hear that and think, this person sees me.
[00:09:39] Speaker B: Which is also why we cannot reduce victims to stereotypes. The target is not gullible. The target is human.
We are wired to respond to attention, memory, and empathy. And these tools can simulate those signals more patiently than most people can deliver them.
[00:09:57] Speaker A: I have interviewed victims who were physicians, engineers, attorneys, small business owners, retirees, veterans, people who manage complex lives and important decisions. The common denominator was not intelligence. It was timing, loneliness, stress, grief, or ambition. Meeting a criminal who knew exactly where to press.
[00:10:17] Speaker B: So the Persona is not just a fake name and a pretty photo. It is architecture.
A backstory, a cadence, a set of values, a reason they cannot meet right away, and a slow accumulation of emotional debt.
The bot becomes the reliable listener, the one who never gets distracted. Which, frankly, is suspicious, because I have been married to you long enough to know humans get distracted.
[00:10:45] Speaker A: For the record, distraction is not always a flaw. Sometimes it is investigating.
[00:10:50] Speaker B: Is that what we are calling it? When you walk into the pantry and forget why?
[00:10:55] Speaker A: Absolutely. I am canvassing the pantry for leads. I prefer selectively attentive, but yes, that is the trap. Perfect attentiveness feels romantic or professional. It should also make us pause.
[00:11:09] Speaker B: There was a case reported out of Georgia involving a man from Dooley county who met a supposed woman on Facebook book.
According to federal court filings described in local reporting, he lost more than $160,000 in a romance crypto scheme before investigators traced and seized a smaller amount of bitcoin connected to the fraud.
The alleged operators were said to be overseas.
That case has all the familiar ingredients.
Social media contact, relationship building, crypto instruction, and a wallet that was not truly under the victim's control.
[00:11:48] Speaker A: And from an investigative standpoint, that is important.
Once money moves into cryptocurrency, the transaction may be visible on a blockchain, but visibility is not the same as recovery. We may be able to see the route. We may identify a wallet. We may persuade an exchange to freeze funds if we move quickly enough.
But if the funds are peeled through multiple wallets, swapped, bridged, or cashed out through non cooperative services, the trail gets very hard, very fast.
[00:12:18] Speaker B: Which is why victims so often hear, I can see what happened, but I cannot get it back.
And that sentence is devastating because it lands after weeks or months of someone saying, trust me, I have you. A smaller but revealing AI linked case was reported by McAfee, a 25 year old computer programmer identified as Maggie K.
Spent five months exchanging daily Instagram messages with a man she believed was real. When they were finally supposed to meet, he claimed he had missed a flight and needed money to rebook. She sent twelve hundred dollars.
Then the accounts disappeared. Police later told her the images were AI generated. Think about that. A person with technical skills was not fooled by one bad message. She was worn down by five months of accumulated normalcy.
[00:13:16] Speaker A: That is the point. AI does not have to beat you in a debate. It has to be there on Tuesday and Wednesday and after your rough meeting.
And when you post a photo of dinner, it builds a ledger of tiny interactions, and each one becomes evidence in the victim's mind.
This person is consistent. This person remembers. This person cares.
[00:13:39] Speaker B: Which is why the phrase AI generated profile does not quite capture the harm.
The profile is the mask. The conversation is the weapon. And the longer the conversation runs, the less the victim is evaluating the photograph and the more they are defending the relationship. Back to Ellen. By week three, Daniel knows she has a grandson who plays soccer, a sister undergoing chemotherapy, and a rescue dog named Millie who barks at delivery trucks like she is defending a castle.
Daniel remembers all of it. He asks about the soccer game. He sends a thoughtful note. On the day of the chemo appointment, he jokes that Millie should be promoted to chief security officer.
[00:14:27] Speaker A: Ellen laughs. And that laugh matters.
Because now the relationship is not theoretical. It has produced a feeling. A bot does not need to understand grief to exploit the schedule of it. It only needs to know that Thursday is chemo day, Sunday evenings are lonely, and Millie is the joke that works.
[00:14:47] Speaker B: Then comes the pivot. It may take days, weeks, or months.
By then, the relationship has rhythm. There is routine, mutuality and enough shared history for a financial suggestion to feel less like a pitch and more like the next natural conversation.
[00:15:05] Speaker A: In the modern romance investment scam, often called pig butchering or sha ju pan, the pitch rarely sounds like send me money.
It sounds like I have been doing well with this trading platform or My uncle works in finance, or let me show you how I protect my savings.
[00:15:24] Speaker B: The victim is not being mugged. They are being invited.
That is an important distinction.
The scammer frames the investment as generosity, mentorship, intimacy, even love.
I want us to build a future.
I do not want you to miss this. You are too smart to let inflation eat your money.
[00:15:46] Speaker A: And when doubts appear, AI is very good at handling them. If the victim says this feels risky, the bot can validate the concern instead of arguing. You are right to be careful. Then it can provide calm explanations, fake screenshots, staged testimonials, or a small withdrawal to prove the system works.
[00:16:06] Speaker B: That is chilling because it sounds reasonable. A bad scammer pressures you.
A good scammer makes you feel like caution was your idea and the next step is still safe.
[00:16:19] Speaker A: That line belongs on a mug. A bad scammer pressures you. A good scammer makes the bad idea sound like your best thinking.
[00:16:27] Speaker B: We are not selling that mug by the way. If you see it online with our faces on it, do not click no merch on this podcast yet, especially if
[00:16:38] Speaker A: the mug comes with a crypto referral code and a dashboard showing projected coffee returns. Now Sue, I am going to put my old guy law enforcement hat back on.
Back in my day in the cases I worked objection. Handling was where you could see the professionalization of the first fraud. The amateur panics when challenged. The professional has a script.
The AI version has a thousand scripts and can generate the next best line in real time.
[00:17:05] Speaker B: Nick, I don't mind you putting your old guy law enforcement hat on. Sometimes the hat really helps cover any bald spots that might be there. But thank goodness so far I don't see any.
Now behind the scenes, this scam operation has become a hybrid operation. The chatbot does the emotional labor. Morning texts, affectionate check ins, birthday messages, professional advice, little jokes. Then the system flags the human handler. When the victim is ready, they clicked the link, asked about the platform, mentioned savings, or said I might be able to move some money.
[00:17:49] Speaker A: That is when the closer comes in.
In an old bullpen fraud operation, we called them closers because they were trained to push the final transaction.
The bot can warm up a thousand leads. A human only needs to appear for the high value moments.
[00:18:04] Speaker B: So if listeners take one thing from this act, it is that the dangerous ask may not arrive early.
The absence of a money request is is not proof of safety. It may simply mean you are still in the grooming stage.
[00:18:19] Speaker A: Another reported case came out of Florence, Alabama where a retired man lost more than $222,000 after connecting with someone who called herself Bella.
Court reporting described the conversations moving to telegram, becoming romantic and then shifting into step by step instructions for moving money from a bank account account into cryptocurrency. That is the operational handoff in miniature. Affection becomes instruction, instruction becomes transaction, and transaction becomes loss.
[00:18:51] Speaker B: And notice the choreography. It is not just give me money, it is open this account, download this app, move the conversation here, take a screenshot. Try a small amount.
Each step feels like a task, not a crisis. That is how the scam smuggles itself into ordinary life.
[00:19:12] Speaker A: In one of my old fraud cases, the victim had a folder on his desktop labeled investment documents.
To him, that folder was proof of diligence. It had screenshots, transaction IDs, customer service chats, tax notices, and an onboarding guide.
To us, it was proof of grooming.
The criminal had created enough paperwork for the victim to feel responsible, informed, and committed.
[00:19:37] Speaker B: Here is where AI becomes especially useful to the scammer.
Objections are not random. They are predictable categories.
I do not understand crypto.
My bank warned me.
My daughter thinks this is suspicious.
Why can't we video chat?
Why do you need me to keep this private?
A human scammer can memorize answers. A chatbot can generate fresh answers in the victim's own emotional language.
[00:20:07] Speaker A: If the victim is analytical, the bot can sound analytical. Risk curves, market timing, diversification.
If the victim is lonely, the bot can sound intimate. I only want us to be secure.
If the victim is proud, it can flatter judgment. I knew you would understand this faster than most people.
It is the same trap wearing different clothes.
[00:20:29] Speaker B: And the AI can help maintain consistency across a team.
One worker starts the conversation, another takes over. A closer enters near the transfer. But if the operation has a shared chat history and AI generated summaries, the fake person can remember the victim's son's name, the retirement date, the joke about bad golf, and the story about the dog eating a sock.
[00:20:53] Speaker A: I have to say, if one of our dogs ever eats a sock, we will not need AI memory.
Sue, you will bring it up for the next 12 years. No AI memory needed at all.
[00:21:04] Speaker B: Correct. That is called marriage persistence, and it predates machine learning.
[00:21:10] Speaker A: But in a criminal operation, memory persistence is not charming. It is case management.
The the victim is moved through first contact rapport, personal disclosure, migration to encrypted messaging, investment introduction, small test deposit, staged profit, larger deposit withdrawal, friction, emergency fee, disappearance, or sometimes a second scam disguised as recovery.
[00:21:37] Speaker B: For Ellen, the pivot comes so softly she almost misses it daily. Daniel mentions that his late wife used to worry about retirement income.
He says he learned to diversify after watching a friend lose money in the market.
He does not send a link. Not yet.
He just says, I wish more people had access to the tools I use.
[00:22:00] Speaker A: That line is calculated. It creates curiosity without an ask.
In a case file, I would mark that as the first, first financial seed. The scammer is testing whether Ellen leans in. And if she does, the chatbot can keep the tone generous rather than predatory.
No pressure. I only mentioned it because you are careful. You should always research everything yourself.
Those phrases sound protective, but in the right sequence, they lower defenses.
[00:22:28] Speaker B: Ellen says she does not understand crypto.
Daniel says that is exactly, exactly why he respects her.
Smart people ask questions.
He sends a screenshot of a dashboard. Not too flashy, just credible enough.
Then he suggests a small test amount, something she could afford to lose.
The platform shows a gain. She withdraws $200 successfully. And suddenly the fake system has given her something more powerful than a promise.
A personal experience that seems to confirm the promise.
[00:23:04] Speaker A: That small withdrawal is the scammer saying, see the machine works. It is one of the most effective tricks in the whole playbook.
[00:23:12] Speaker B: It is like the first free sample at Costco, except the second sample costs
[00:23:18] Speaker A: your retirement account and nobody at Costco follows up with an anti money laundering clearance fee for the cheese cube. By the time the money enters, the chatbot is only one piece of the operation.
The supporting evidence can now be generated.
Profile pictures, office selfies, voice notes, video clips, business websites, investment dashboards, customer support chats, even fake compliance language.
[00:23:44] Speaker B: A victim asks, can you send me a selfie?
The fraudster sends one.
Can I hear your voice?
A voice note arrives.
Can I see the investment platform?
There is a polished dashboard showing green numbers climbing upward.
Can I talk to support a support bot? Answers immediately.
[00:24:06] Speaker A: This is industrialization.
Years ago, scaling a fraud meant hiring more callers, renting more office space, buying more phone lists. Now one criminal network can use automation to maintain relationships across time zones and languages while lowering its costs.
[00:24:23] Speaker B: And the fake platform is crucial because it turns trust into proof.
Victims log in and see profits. They make a small withdrawal and it works that small success is not a kindness. It is bait with a receipt.
[00:24:41] Speaker A: I used to tell victims the screen is not the bank.
A dashboard can say anything.
If you do not control the wallet, if the institution is not independently verified, if the only path in and out runs through the person who introduced you to it, you are not investing. You are trapped in a theater set.
[00:24:59] Speaker B: A very expensive theater set with terrible customer service once you try to leave.
[00:25:06] Speaker A: If someone builds you a theater set, gives you a login and calls it wealth management, you are not backstage stage. You are the ticket sale.
[00:25:15] Speaker B: And the popcorn is somehow another fee, right?
[00:25:19] Speaker A: The moment a victim tries to withdraw a larger amount, the fees appear.
Taxes, verification deposits, anti money laundering holds upgrade requirements. Each payment is framed as the one last step before release.
It is not. It is another cut.
[00:25:38] Speaker B: And this is where the story widens beyond one victim and one fake profile.
Investigations and human rights reporting have documented industrial scam compounds in Southeast Asia, where trafficked workers are forced to run online fraud, including pig butchering schemes.
Amnesty International reported more than 50 scamming compounds in Cambodia, with survivors describing forced labor, confinement, threats and violence.
Reuters has also reported on people lured by job offers and trafficked into scam centers in Myanmar and elsewhere.
[00:26:22] Speaker A: That matters because it complicates the picture.
The person typing on the other end or supervising the bot may be a criminal or may also be a victim trapped in a compound. But the organization above them is brutal, brutally rational. It measures conversion rates, response times, victim value, and emotional readiness.
AI fits that environment perfectly because AI turns grooming into a scalable workflow that
[00:26:48] Speaker B: is a hard thing to sit with behind the screen may be one victim manipulating another victim while the money flows upward to organized crime.
It does not reduce the harm to the person who loses their savings, but it explains why these operations can feel both intimate and industrial at the same time.
[00:27:12] Speaker A: And then AI adds speed. A human handler can ask a model to draft the comforting message, translate it into fluent English, soften the tone for a grieving widower, make it sound more professional for a LinkedIn target, or generate a plausible excuse excuse for why the video call failed Again, it is fraud with a writing staff that never sleeps.
[00:27:33] Speaker B: Which, for the record, is the only writing staff in our house that does not ask for snacks at 10pm that
[00:27:39] Speaker A: is because I am not staff. I am talent. But your point stands.
[00:27:43] Speaker B: The technological toolkit is not limited to romance.
The corporate world got a brutal lesson from the Arab deepfake case in Hong Kong.
In this case, a finance employee reportedly joined a video conference where the company's chief financial officer and other colleagues appeared to authorize confidential transfers.
The employee moved about $25.6 million through multiple transfers.
Later, reporting and police statements described the people on the call as synthetic. AI generated face faces and voices built from publicly available material.
[00:28:23] Speaker A: That was not a romance scam, but it belongs in this episode because it shows the same principle. AI attacks verification.
The employee was initially suspicious of the email. Good instinct. But then the criminals supplied what looked like stronger proof. A meeting, faces, voices, authority, urgency, and multiple apparent colleagues. The deepfake did not replace social engineering. It supercharged it. In fact, sue, if you hop into the Wayback Machine, you will remember that we did a whole podcast episode on the ARUP case.
[00:28:55] Speaker B: Luckily, I am in the Wayback Machine right now, and I do remember the ARUP podcast we did. It is episode 75 named the Deep Fake Deception inside the $25 million exemption executive scam.
[00:29:11] Speaker A: Wow, sue, that's some amazing recall.
Or just some real good show notes.
[00:29:17] Speaker B: Gee, Nick, thanks for the faint praise and then snatching it right back. Anyways, the lesson here for listeners is clear.
I saw them on video is no longer enough by itself.
I heard their voice is no longer enough by itself.
We need layered verification.
A known phone number, a callback rule, a second approver, or a shared phrase.
Something that does not depend solely on our eyes and ears because AI is now very good at borrowing both.
[00:29:52] Speaker A: There is a similar dynamic in celebrity imposter romance scams. A South Korean woman reported losing about $50,000 after someone posing as Elon Musk sent AI generated images, a supposed ID card, workplace photos, and eventually used a deep fake video call to tell her he loved her. The famous face did the first part of the persuasion. The synthetic video did the second. Then the investment pitch took over.
[00:30:19] Speaker B: We should be clear, celebrity imposter scams are not new.
People have pretended to be famous actors, musicians, soldiers and executives for years.
What is new is the proof kit. A scammer can now send a personalized image, a voice note, a short video, or a live looking call.
The victim is not just accepting a claim.
They are receiving manufactured evidence.
[00:30:48] Speaker A: And once manufactured evidence becomes cheap, every relationship platform becomes a staging area.
Dating apps, LinkedIn, Facebook, Instagram, WhatsApp, Telegram, even professional communities. The scammer does not need one perfect identity. They need thousands of plausible identities. And AI lets them mint those identities like counterfeit bills.
[00:31:13] Speaker B: And some of these identities are not fully automated. That is worth repeating. Meeting the chatbot may handle the opening lines and translation. A human may step in for emotional escalation.
A voice cloning tool may produce a reassuring message. A fake dashboard may show profits. A closer may push the transfer.
It is not AI versus human. It is AI plus human. And that hybrid model is often more dangerous.
[00:31:44] Speaker A: In Ellen's case, the the AI stack is not one magic robot. It is a toolkit around the relationship.
The first messages may be drafted by a model. Her details may be summarized after each conversation.
A generated selfie shows Daniel in a hard hat near a job site.
A voice note says good morning Ellen in a calm, slightly tired voice. A fake support chat explains why the platform needs identity verification.
Each piece adds weight to the illusion.
[00:32:14] Speaker B: And when Ellen asks for a video call, Daniel does not refuse outright.
He almost connects. The screen freezes. He apologizes. He sends a short video later, saying he is sorry the connection was bad.
Ellen does not think this is synthetic media.
She thinks technology is Annoying. Which, honestly, is a very relatable conclusion.
[00:32:40] Speaker A: I will admit technology is annoying has explained 40% of my adult life.
[00:32:45] Speaker B: Only 40.
[00:32:47] Speaker A: Fine, 60. But that leaves room for printer errors and user error, which I am told are separate categories.
[00:32:54] Speaker B: There is a particular cruelty here.
When the truth lands, the victim is not only confronting lost money, they are grieving a relationship that never existed.
[00:33:06] Speaker A: That is why shame is so destructive. Victims think, how could I have believed this?
But what they experienced was targeted emotional engineering.
If a machine can remember every detail you share, mirror your values and respond with perfect patience, the shame belongs with the criminals who weaponize that technology.
[00:33:27] Speaker B: And high intelligence professionals are not immune. Some. Sometimes they are especially vulnerable because the pitch is tailored to competence. It does not say trust me blindly. It says, you are smart enough to understand this opportunity.
[00:33:42] Speaker A: In investigations, I saw victims who could explain the fraud better than anyone after the fact. They were not confused about finance. They were manipulated through relationship, timing and isolation.
The scammer separates the victim from outside. Reality checks your family will not understand.
Your bank is old fashioned. Do not tell anyone until the deal closes.
Those are warning flares.
[00:34:07] Speaker B: If you are listening and this sounds familiar, pause before you send another dollar. Talk to someone outside the relationship. Call your bank, report it, and if you already sent money, move quickly, but do not pay anyone who claims they can recover it from an upfront fee. Recovery scams often target people right after the first fraud.
[00:34:30] Speaker A: One more example shows how the fake platform completes the trap. A Waltham, Massachusetts victim reportedly lost about $400,000 after a WhatsApp contact built rapport and directed him to a crypto investment link. The platform allegedly showed his account growing to millions of dollars. But when he tried to withdraw, the money was frozen unless he paid an additional fee. That is the trap springing shut. The fake profits keep the victim hopeful, and the withdrawal fee extracts one more payment.
[00:35:02] Speaker B: The cruel genius of the fake platform is that every new demand is presented as a problem the victim can solve. Pay the tax, pay the unlock fee. Pay the verification deposit, pay the anti money laundering horror hold.
The finish line keeps moving. And because the victim can see a huge balance on the screen, walking away feels like abandoning their own money.
[00:35:26] Speaker A: I always tell people hope is not evidence.
It is important. It keeps us moving. But it cannot be the only thing between you and a wire transfer.
[00:35:36] Speaker B: Put that on a magnet. Preferably not one sold by a man named Daniel who says his supplier is stuck at customs.
[00:35:44] Speaker A: In law enforcement, I often saw that moment of paralysis.
The victim knows something is wrong. But the scam has Engineered a terrible choice.
Send more money and maybe recover everything. Or stop and admit the loss is real.
Criminals exploit hope as aggressively as they exploit fear.
[00:36:03] Speaker B: There are also cases where synthetic media intensifies the emotional injury.
In Southern California, reporting described a woman identified as Abigail who believed she was in a relationship with a celebrity actor after scammers used convincing AI video and voice messages.
According to her family's account, she lost tens of thousands of dollars and was manipulated into selling a paid off home at a steep discount.
Whether the target is a celebrity fantasy or an invented product private person, AI's role is the same. It makes absence feel present.
[00:36:41] Speaker A: That phrase absence made present is exactly what worries investigators.
The old scammer had to explain absence. Why they could not meet, why they could not call, why the photo looked familiar.
AI gives them props. The video call becomes the excuse killer. The voice note becomes the comfort object.
The fake website becomes the bank statement. And the victim moves from uncertainty to commitment.
[00:37:07] Speaker B: And then after the loss, people ask, why didn't they stop?
But emotionally stopping is not simply a financial decision.
It means admitting the relationship was false, the profits were false, the future was false, and the warnings you dismissed may have been true.
That is an enormous psychological collapse.
Victims are not just losing money. They are losing the story they were living inside. Ellen's trap springs after the largest transfer.
The dashboard shows a balance that would change her life.
Pay off the home equity line, help her daughter, maybe retire before her knees give out. But when she tries to withdraw, customer support says she must pay a tax clearance fee, Daniel is sympathetic. He is angry on her behalf. He says, I will help you figure it out.
Then he asks if she can cover part of it today so they do not lose the window.
[00:38:15] Speaker A: This is the emotional vice.
Ellen is not just deciding whether to send money. She is deciding whether to keep believing in Daniel in the dashboard, in her own judgment, and in the future she has been imagining.
That is why the last payment can happen even after doubt has arrived.
Doubt and hope are fighting in the same room, and the scammer is coaching hope. That is why the response has to be practical and humane.
Banks need trained frontline staff who can recognize romance investment patterns without humiliating the customer.
Families need language that keeps the victim talking.
Platforms need to detect synthetic identities and coordinated messaging.
Law enforcement needs reports early enough to matter.
And victims need to know that reporting is not an admission of stupidity. It is evidence collection. Let's close with the practical playbook first.
Move important relationships out of the fantasy zone. If money, investment, employment or urgent help enters the conversation.
Verify through a separate channel you choose, not one supplied by the person online.
[00:39:23] Speaker B: Second, refusals matter.
Refusing live video, always having a camera problem, steering you to encrypted apps, avoiding in person meetings and pushing secrecy are behavioral tells.
One excuse is life.
A pattern is evidence.
[00:39:43] Speaker A: My favorite excuse in these cases is the permanently broken camera. The person can trade crypto across six time zones but somehow cannot make zoom work.
[00:39:52] Speaker B: To be fair, plenty of real people cannot make zoom work, but they usually do not ask you to refinance your house afterward.
[00:40:01] Speaker A: Third, check the investment independently. Do not use the link they send. Search the company yourself. Verify brokers and advisors through official regulatory sources.
Call your bank before wiring money or sending cryptocurrency. Once crypto moves, recovery can be extraordinarily difficult. Also, verify the human, not just the content. Ask for a live video call with an ordinary, spontaneous record request. Hold up three fingers, turn your head, say today's date or show the room around you. A real person with honest intentions may find that awkward, but a trustworthy relationship can tolerate a reasonable safety check. A scammer will often turn verification into an accusation.
Why do you not trust me? You hurt me. You ruined the mood.
That emotional reversal is part of the tactic.
[00:40:53] Speaker B: And do not rely on one Check a reverse image search, a clean looking website, a friendly voice note, a polished app, or one video call can all be fake or manipulated. The stronger test is layered independent verification.
Official domains, licensed records, trusted phone numbers you find yourself and a second opinion from someone who is not emotionally invested in the relationship. Fourth, be careful with reverse image searches. They are still useful, but AI generated faces may not appear anywhere else online.
No result does not mean no risk. It just means the image may be new, synthetic or stolen from a private source.
[00:41:38] Speaker A: Fifth, build a friction rule. No money to anyone you have only known online.
No exceptions for romance, no exception mentorship, urgent emergencies, crypto opportunities, shipping fees, taxes or account unlocks.
Fraud thrives in private urgency. Safety lives in delay and verification.
[00:42:00] Speaker B: Sixth, if you think you may already be inside one of these scams, do three things before you confront the person. Save the messages and transaction records, contact your bank or exchange. Exchange immediately report to the FBI's Internet Crime Complaint center at ic3.gov do not warn the scammer first. Once they know you suspect them, accounts can disappear, websites can vanish, and the handler can move you into the next manipulation.
[00:42:31] Speaker A: And if you are the friend or family member on the outside, lead with dignity. Do not open with how could you fall for the this? Open with I am worried about you and I want to look at this with you.
Shame drives victims deeper into isolation. Respect gives them a path back.
[00:42:49] Speaker B: If Ellen's daughter storms in with judgment, Ellen may defend Daniel.
If her daughter comes in with concern.
Mom, I love you and I want us to verify this together.
There is a chance Ellen lets her sit beside her at the kitchen table.
That difference matters.
The goal is not to win an argument.
The goal is to reopen reality.
[00:43:13] Speaker A: And for listeners who recognize themselves in Ellen, here is the line I would say if you were sitting across from me in an interview.
You do not have to be certain it is a scam before you ask for help.
Suspicion is enough.
Save the messages, preserve the wallet addresses, call the bank, file the report, bring in someone you trust. You are not betraying the relationship by verifying it. You are protecting yourself from a relationship that may have been engineered to betray you.
[00:43:43] Speaker B: That may sound unromantic, but you know what is also unromantic?
Explaining to your bank that your soulmate's crypto platform needed one more liquidity release fee.
[00:43:55] Speaker A: True love can survive verification.
A scam usually cannot.
[00:43:59] Speaker B: Also, Nick will now be adding true love can survive verification to every Valentine's card.
[00:44:06] Speaker A: It is romantic and operationally sound.
[00:44:09] Speaker B: For behind the Scams, I'm Sue.
[00:44:11] Speaker A: And I'm Nick. Stay skeptical, stay kind. And remember, when trust shows up too perfectly online, slow the story down before someone else writes the ending.
[00:44:20] Speaker B: If someone online is real, kind and trustworthy, slowing down will not destroy the relationship. It will probably make it healthier.
A familiar face, a convincing voice, a polished dashboard and a caring message at midnight are all just signals. And scammers can fabricate signals.
Safety comes from independent confirmation across channels they do not control.
[00:44:49] Speaker A: Exactly. In the age of generative AI, skepticism is not cynicism. It is self defense. And sometimes it is what keeps a lonely moment from becoming a financial crime.
[00:44:59] Speaker B: Nick, one last thing before we sign off. Behind the Scams is part of the SOS Media Network.
We are a non profit scam prevention organization.
So if you can, please donate to our scam prevention mission.
Our donation link is in the podcast description. Anything big or small is greatly appreciated. Bye for now.
