AI Romance Scams: How Deepfakes and Chatbots Supercharged Catfishing

---

Americans lost more than $1.14 billion to romance scams in a single year, with a median loss of $2,000 per person, the highest of any imposter scam the FTC tracks, per the FTC Consumer Sentinel Network Data Book. Those numbers came before AI made the fakes nearly perfect. Today the old advice, "just ask for a video call," no longer protects you.

What are AI romance scams? AI romance scams use generated profile photos, deepfake video calls, and chatbots to run convincing fake relationships at scale. Old safety checks like requesting a live video call can now be defeated. Verify a match with reverse image search, an AI image detector, unpredictable live-call requests, and a firm refusal to ever send money.

If you are reading this because something feels off, trust that instinct. The people who fall for these scams are not careless. The FTC notes victims are often educated and cautious. The difference now is that the criminals run an industrialized operation, and AI hands them tools that defeat the checks people were taught to rely on.

How AI Changed Romance Scams

Romance scams are not new. What changed is the tooling. A scammer used to need stolen photos, decent English, and a plausible excuse for never appearing on camera. Each of those was a weak point you could catch. Generative AI removed all three weak points at once.

The result is faster, cheaper, and far harder to detect by eye. A single operator can now run dozens of "relationships" in parallel, each with a unique face that exists nowhere else, backed by a chatbot that never sleeps and never forgets the script.

The financial endgame has also industrialized. The pig butchering scam, named for the way victims are "fattened up" with affection before the slaughter, blends romance grooming with fake crypto investment platforms run by organized crime. The FBI's IC3 tracks this under confidence and romance fraud, which accounted for roughly $929 million in reported losses in 2024, per the FBI IC3 2024 Internet Crime Report.

The 3 AI Upgrades Scammers Use

Each AI upgrade defeats a safety check that used to work. Here is how each one works, and the countermeasure that still holds.

AI-generated profile photos that beat reverse image search

For years the standard advice was to reverse image search a match's photos. If the same face turned up on a stranger's account or a stock site, you had your answer. That check assumed the photo was stolen from a real person.

Generative tools like those behind "this person does not exist" sites produce faces that belong to no one. There is no original to find, so reverse image search comes back empty, which can falsely reassure you. This is exactly why an AI image detector matters now: it looks for the synthetic fingerprints in the image itself rather than searching for a match elsewhere. Telltale artifacts still show up on close inspection, such as mismatched earrings, melted jewelry, garbled text on signs, irregular teeth, and backgrounds that warp near the edges.

Deepfake video calls: why "let's video chat" is no longer proof

Insisting on a video call was the gold-standard test. A scammer using stolen photos could not show up live, so the request usually ended the relationship. Real-time face swap broke that test. Attackers now run consumer-grade software that maps a synthetic or stolen face onto their own in real time during a call.

Quality varies, and that variation is your opening. Deepfake video calls often look slightly off under stress: the face flickers at the edges, lighting on the face does not match the room, blinking looks unnatural, and the picture stays oddly low-resolution to hide artifacts. The countermeasure is to make unpredictable requests the software cannot render cleanly, covered in the verification section below. For more on this specific attack, see our guide to deepfake video call scams.

Chatbot grooming: one scammer, hundreds of victims

The most invisible upgrade is conversational. Large language models let one operator groom hundreds of victims at once, each conversation fluent, attentive, and emotionally tuned. The grammar mistakes that used to give scammers away are gone. Love bombing, the flood of intense affection early on, scales effortlessly when a machine writes it.

The weak point is memory and consistency. A chatbot juggling many targets, or an operator copying replies between conversations, will contradict itself: forgetting details you shared, mixing up your job or city, or repeating a story with different facts. Genuine partners remember. Watch for a relationship that feels intensely attentive yet strangely impersonal under the surface.

10 Warning Signs of an AI Romance Scam

No single sign is proof. Several together are a strong signal to slow down and verify.

1. Photos look too perfect. Polished, model-grade images, or photos with subtle artifacts on ears, teeth, hands, or jewelry.
2. Fast, intense love bombing. Declarations of love within days, talk of a future before you have met.
3. They never meet in person. Every in-person plan collapses with a last-minute emergency.
4. Video calls are short, glitchy, or low light. The image is poor, the face flickers, or they cut calls quickly.
5. Inconsistent memory. They forget what you told them, or facts about their own life shift over time.
6. They push you off the dating app. Quick moves to WhatsApp, Telegram, or text, away from platform moderation.
7. Money or crypto enters the conversation. Any direction of money talk is a red flag, especially early.
8. An investment "opportunity." They are making great returns on a crypto or trading platform and want to help you do the same. This is pig butchering.
9. Emergencies with deadlines. A medical crisis, a stuck shipment, a customs fee, always urgent, always needing money now.
10. Requests for gift cards or transfers. Gift cards, wire transfers, crypto, or "holding" money for them. Legitimate partners do not ask.

For a deeper look at synthetic profile images specifically, see our guide to fake profile picture detection.

How to Verify an Online Match Is a Real Person

You do not need to be a forensic analyst. A short verification routine catches the overwhelming majority of AI romance scams.

1. Reverse image search first. Still worth doing. If their photo turns up on a stranger's profile or a stock site, you are done. If it finds nothing, that is not proof of a real person, because generated faces have no original. Move to step two.
2. Run the photos through an AI image detector. Because reverse image search fails on generated faces, analyze the image itself for synthetic fingerprints. Save a profile photo and upload it to a detector for a verdict.
3. Make an unpredictable live-call request. On a video call, ask them to turn their head fully to the side, wave a hand slowly in front of their face, or hold up a specific everyday object and move it. Real-time face swaps struggle with sharp angles, occlusion, and fast motion. A real person does it without hesitation.
4. Meet in public when local. If they claim to be near you, suggest a quick coffee in a public place. A persona that does not exist will always have a reason it cannot happen.
5. Never send money, regardless. No verification result changes this rule. Do not send money, gift cards, or crypto, and do not "invest" through a platform a match introduced you to. There is no scenario where this is safe.

CTA: Unsure about a match's photos or videos? Check them free, 50 detections per month, files purged within 60 seconds of analysis. Create a free account

When you run a check, our detector returns a clear verdict, Authentic, Likely Synthetic, or Inconclusive, with a TrustScore from 0 to 100, so you get a confidence read rather than a coin flip. It analyzes images, video, and audio, and supported image files are JPG, PNG, or WEBP up to 4.5MB.

If You Already Sent Money to a Romance Scammer

If this already happened to you, you are not foolish, and you are not alone. This is an organized crime, run by people who do this professionally, full time, at scale. Act fast and be kind to yourself.

1. Contact your bank or payment provider immediately. Speed matters most for recalls. Ask about reversing transfers and freezing further activity. For crypto, contact the exchange you used.
2. Stop all contact and preserve everything. Do not warn the scammer. Screenshot chats, profiles, photos, payment records, and account names before anything disappears.
3. Report it. File with the FTC at reportfraud.ftc.gov and with the FBI's Internet Crime Complaint Center at ic3.gov. Report the profile to the dating app or platform as well.
4. Beware recovery scams. Criminals target recent victims with offers to "recover" lost funds for a fee. Legitimate authorities never charge you to get your money back. Treat any such offer as a second scam.

For a full step-by-step, see our deepfake scam response playbook. If you are worried about a parent or older relative, approach with empathy, not accusation. Victims often defend the scammer, and shame keeps them silent. Lead with concern, not "I told you so."

FAQ

How much do people lose to romance scams? A lot, and it is rising with AI. The FTC's Consumer Sentinel Data Book reported about $1.14 billion in romance scam losses with a median loss of $2,000 per person, the highest of any imposter scam, per the FTC. The FBI's IC3 separately reported roughly $929 million in confidence and romance fraud losses in 2024.

Can scammers fake a live video call? Yes. Real-time face swap software runs on ordinary consumer hardware, so "let's video chat" is no longer proof of a real person. Quality drops under stress, so unpredictable physical requests, like turning the head fully sideways or waving a hand across the face, still break most fakes.

How can I tell if a profile photo is AI generated? Reverse image search will not help, because generated faces have no original to find. Run the photo through an AI image detector that analyzes the image itself, and inspect the details: mismatched earrings, distorted jewelry, irregular teeth, garbled background text, and warped edges are common giveaways.

Why do romance scammers avoid meeting in person? Because the person you are talking to does not exist. The photos are generated or stolen, and the chats may be written by a bot. Every in-person meeting gets an endless string of emergencies and excuses, because there is no real person who can show up.

Who do I tell if my parent is caught in a romance scam? Report to the FTC at reportfraud.ftc.gov, the FBI's IC3 at ic3.gov, and your parent's bank, which can flag or freeze suspicious transfers. Approach your parent with empathy. Victims often defend the scammer and feel deep shame, so accusations push them away. Focus on stopping further losses.

Conclusion: Love Should Survive Verification

AI romance scams work by making the fake feel real: generated faces that beat reverse image search, deepfake video calls that defeat "let's chat live," and chatbots that love-bomb at scale. The defense is not cynicism, it is a simple routine. Reverse image search, an AI image detector, an unpredictable live-call request, and an unbreakable rule never to send money. A real person who cares about you will tolerate a few verification steps. A scam will not survive them. If a match's photos or video feel off, check them before your heart, or your savings, is on the line.

CTA: Verify a match's photos or video free, 50 detections per month, files deleted within 60 seconds of analysis. Get started free

Related reading