The risks of online dating haven't gone away — they've gotten more sophisticated. AI-generated profiles, pig butchering scams, and account cloning are the new normal. Here's how to navigate all of it.
Over 300 million people use dating apps globally. The FTC reported $700 million in romance scam losses in 2024 — and that number only covers cases that were reported. The real figure is higher. Most people who get defrauded don't tell anyone, because the embarrassment feels too hard.
None of this means online dating is a mistake. It means you should go in with your eyes open. The people who get hurt most are usually the ones who thought they were too smart to be fooled — not because they were overconfident, but because they didn't know what the current threats actually look like. This guide covers all of it.
The threat landscape has shifted. A few years ago, the main risk was someone using a stolen photo from a real person's Instagram account — and reverse image search could catch it. That approach has been largely superseded.
AI-generated profile photos are now the primary tool for large-scale fake profile operations. An AI face can be generated in seconds for free. It doesn't exist anywhere else on the internet, which means reverse image search finds nothing. The face looks real — often more than real, with a specific too-polished quality that most people don't consciously notice. Sumsub's 2024 report found that deepfake-related identity fraud grew 257% in a single year.
Pig butchering is the romance scam variant that's produced the largest losses. It starts like a normal relationship — weeks or months of real conversation and emotional connection — and ends with an introduction to a fake cryptocurrency investment platform. By the time the financial ask arrives, the relationship feels real enough that the investment seems reasonable. FBI and CISA have both issued formal warnings. Losses in the US alone run into the hundreds of millions annually.
Profile cloning is still common on Facebook especially — someone creates an account that mirrors a real person's profile, then sends friend requests to all their contacts. The clone requests personal information, money, or both. The original person often doesn't find out for days.
Sextortion has increased sharply, particularly targeting younger users. Someone sends intimate photos and immediately requests the same in return — then uses those photos to extort payment. The initial photos are often AI-generated or recycled from previous victims.
Each platform has its own risk profile. They all have fake accounts — the specific type of risk varies.
Tinder has photo verification for some accounts, but it doesn't verify that profile photos match the verified person. AI-generated main photos attached to verified accounts are possible and happen. Volume is high, which makes individual moderation difficult.
Bumble has a Deception Detector that catches behavioral patterns — bot-like activity, copy-paste messaging. It doesn't analyze photos for AI generation. Women-message-first doesn't eliminate fake accounts; it just changes which direction the first message comes from.
Hinge positions itself as relationship-oriented, which makes users more likely to invest emotionally early. That investment is exactly what a scammer wants. "Designed to be deleted" creates a sense of seriousness that makes a manipulative actor seem more trustworthy.
Facebook has the highest fake account volume of any platform — Meta's own transparency reports put it in the hundreds of millions per quarter. Profile cloning and Marketplace fraud are both common. Facebook Dating within the app inherits all of these problems.
Instagram is a common target for crypto and investment scam approaches — someone follows you, engages with your posts for days, then introduces an "opportunity." The engagement feels genuine before the pitch arrives.
LinkedIn removed 95 million fake accounts in 2023. Fake recruiters, phantom hiring managers, and competitor intelligence accounts all use AI-generated headshots because they look professional. Pig butchering operations have also moved onto LinkedIn because users there often have more disposable income.
The best time to catch a fake profile is before any conversation starts. Once you've been talking to someone for a week, the emotional math changes. A red flag you notice on day one is easy to act on. The same red flag on day 10 feels like an accusation against someone you've gotten to like.
Before swiping right or accepting a connection, take 60 seconds:
Run the profile photo through Faux Spy. An AI-generated face comes back as "AI Photo" with a confidence score. Above 80% is a strong signal. Below 50% probably means the image is a real photograph.
Then run a reverse image search — right-click the photo and select "Search image with Google," or upload to TinEye. If the photo appears on someone else's profile, you're looking at a cloned account.
The two checks together take about a minute and cover both main methods of fake profile photos. A fake profile using an AI-generated image can survive the reverse image search. A real but stolen photo gets caught by the reverse search but not necessarily by the AI detector. You need both.
Some red flags show up early. Others don't appear until weeks in. Here's the pattern in rough order of when they tend to surface.
Pressure to move off-app immediately. "I'm not on here much — can we text?" This is the most consistent behavior across all romance scam types. Dating apps have moderation. External messaging apps don't. If someone is pushing for WhatsApp or Telegram within the first few messages, that's worth noting.
Emotional intensity that outpaces the timeline. Real connection takes time. A script designed to manufacture connection moves fast. Strong feelings, future plans, "you're different from anyone I've met" — before you've had a video call — is the script, not spontaneous feeling.
Video calls that never happen. Bad connection. Broken camera. Not comfortable yet. Each excuse sounds plausible once. Together, over weeks, they form a pattern that real people don't have. Someone who likes you will find a way to be on camera. A fake account literally cannot.
A backstory that explains unavailability. Military deployment. International work contract. Oil rig. These are chosen because they sound legitimate and conveniently explain why meeting in person is off the table indefinitely.
For a complete list, see the guide on romance scam warning signs and the blog post on 10 signs someone is catfishing you.
If you've been talking to someone for weeks and want to meet in person, do one thing first: a live video call with a specific challenge.
Don't just ask for video — ask them to do something specific in real time. "Can you wave at me with your right hand?" or "Can you hold up your fingers so I can see how many?" or "Can you show me what's out your window right now?" A pre-recorded video can't respond to a real-time request. An AI-generated video feed has limited ability to follow live instructions.
If they've been unable to video call for weeks and still can't manage it before an in-person meeting, that's a decisive data point.
When you do meet: choose a public location for the first meeting. Tell a friend or family member where you're going and who you're meeting. Share your location with someone you trust. This isn't paranoia — it's just reasonable caution for a first meeting with a stranger from the internet.
Don't send money to anyone you haven't met in person. Not a loan. Not a gift card. Not a wire transfer. Not crypto. Not "holding" money for them.
This rule sounds obvious. It isn't, once you're inside a relationship that feels real. The ask doesn't come until after weeks of genuine-feeling connection. The story is sympathetic — a medical emergency, a stuck customs payment, a once-in-a-lifetime investment. The emotional architecture has been carefully built to make the request feel reasonable.
The rule isn't "be suspicious of everyone." It's simpler than that: if you haven't met them in person, don't send them money. Full stop. Not even once. Not even a small amount to test the waters.
A real person who cares about you won't ask for money before you've met. If they do, that's the answer to the question about whether this is real.
If money has been sent: stop immediately. Don't pay any additional fees, taxes, or "compliance charges" to release your funds — those are secondary scams designed to extract more money from the same victim. The original money is almost certainly gone.
Report to the FTC at ReportFraud.ftc.gov. Report to the FBI's IC3 at ic3.gov. Contact your bank about any wire transfers or ACH payments as quickly as possible — some can still be reversed if caught early enough.
If intimate images were shared and are being used for extortion: contact the FBI's IC3 and the platform where the contact occurred. The StopNCII database allows you to register images to prevent them from spreading. Do not pay — payment rarely stops the extortion and often escalates it.
Don't be embarrassed into not reporting. The FTC data on romance scam losses is almost certainly lower than reality because most victims don't come forward. Every report helps agencies build cases against operations that are still actively running.
Estimates vary, but independent researchers consistently find that 10-20% of profiles on major dating apps include some form of deception. AI-generated profile photos now make up the majority of fake profile photos because they can't be caught by reverse image search.
Pig butchering is a romance scam that incorporates a fake cryptocurrency investment platform. After weeks of building a real-feeling relationship, the scammer introduces an investment opportunity with guaranteed returns. Victims invest money, see apparent gains on a fake dashboard, invest more, and then discover that all fees and "release payments" are additional extractions. There are no funds — the platform is entirely fake. The FBI and CISA have both issued formal warnings.
Reverse image search catches stolen real photos — images taken from someone else's account. It doesn't catch AI-generated faces, which are unique images that have never appeared anywhere before. You need both: a deepfake detector for generated faces, and reverse image search for stolen ones. They're complementary, not interchangeable.
Request a live video call with a specific real-time challenge. Check their profile photos with a deepfake detector and reverse image search. Tell a friend where you're going. Meet in a public place. Don't share your home address until you've met in person multiple times.
10 free checks per day. Works on Tinder, Bumble, Hinge, Facebook, Instagram — any website in Chrome.
🕵️ Add to Chrome — Free 🦊 Add to Firefox — Free