Trust & safety · 7 min read
In 2026, spotting a catfish is harder than it used to be. AI-generated photos are convincing, scammers are patient, and "verified" badges have come to mean different things on different apps. Here's a practical field guide.
Join the open beta →Founding $9.99/mo · locked for life · ends Sept 1, 2026Catfishing has changed. Five years ago the giveaways were obvious: stolen modeling photos, broken English, an "oil rig job" or "deployed to Syria" story. Modern romance scams are quieter. They invest weeks before asking for anything. They use AI-generated profile photos that don't exist anywhere else online — so reverse image search returns nothing — and they're often run as small businesses, not lone individuals.
The good news: most catfish profiles still leave a trail of small inconsistencies, and the right verification can catch them before you ever see them.
Suggest a five-minute FaceTime or video call early. A real person who's interested will say yes within the first week or two of texting; a scammer will have an indefinite reason it's not possible. Note that the bar is "video," not "phone" — voice is now AI-fakeable in real time, video is much harder.
Real people accumulate photos over years. Different lighting, different haircuts, different friends, different phone cameras. Scam profiles often look like a curated mood board — six photos that all feel like they were taken in one week, by the same photographer, in the same lighting. Trust your eye on this.
Pay attention to small mentioned facts — the dog's name, the job, where they grew up. Romance scams typically don't keep tight notes on what they've told whom. If the dog has changed names by the third week, you have your answer.
A real Austin person, talking unprompted, will mention specific neighborhoods, traffic, the weather, last week's events, a coffee shop they went to. Scammers tend to stay generic. If you're three weeks in and you haven't heard a single Austin-specific reference that wasn't in response to a direct prompt, be skeptical.
The end-state of a romance scam is always money, but the request is rarely direct. It comes wrapped in a story — a sick relative, a temporarily frozen account, a customs fee, a crypto opportunity they want to share. The rule is simple and absolute: if you have not met someone in person, you do not send them money. Ever. There is no exception.
"Photo verification" means different things on different platforms. The strongest version is a multi-step process: the user records a short selfie video performing specific motions, AI compares it frame-by-frame to their profile photos to confirm the same face, and the resulting hash is checked against every other account on the platform to detect duplicate uses.
On Heart of Gold, every account has to pass this verification before it becomes visible. The selfie is matched to the profile photos via face-recognition AI, and the resulting biometric hash is checked against the rest of the platform — meaning a scammer can't just create twenty accounts with the same photos. If the same face shows up on a second account, both get flagged.
This catches the easy 95%. The remaining 5% — patient, well-resourced scams that pass verification with a real human accomplice — get caught by behavior monitoring later. But verification at the door is what makes the rest of the system tractable.
Want the details? See how Heart of Gold's photo verification and trust system work →
A new wrinkle: profiles where the photos are real (a real person you can verify), but the messages are written by AI. This isn't strictly catfishing — there's a real person on the other end — but it's adjacent. You're falling for a writing voice that isn't theirs.
Tells: messages that are unusually well-structured, the same flowery phrasing every time, an inability to be brief or casual, perfect grammar even at 2am, conversational depth that's slightly off the conversational beat. If you ask a quick clarifying follow-up that requires real personal context, AI-generated replies often miss in a specific way — they answer the previous message rather than the actual current question.
If you're reading this because something already happened — you sent money, you sent photos, the person you thought you knew turned out to be someone else — first, you're not alone, and second, you're not stupid. Romance scams are designed by professionals to bypass exactly the parts of your brain that would normally catch them. The patience, the parallel narratives, the slow build of trust — none of that is amateur work.
Concrete steps: report the account on the platform where it happened. File a report at IC3.gov (the FBI's Internet Crime Complaint Center) — even small dollar amounts matter for triangulation. If you sent money via wire or crypto, contact the institution within 48 hours; recovery rates are low but not zero, and they get worse with time. If photos were involved, organizations like the Cyber Civil Rights Initiative offer free help with takedowns. None of this is fun. All of it is worth doing.
If something feels off, something is off. The number-one regret in every romance-scam survivor account is some version of "I felt like something was wrong, and I talked myself out of it." The patterns described above are useful, but the deepest signal is the one your gut is already giving you.
A useful test: tell one friend — a real friend, not someone in the conversation — about the person you're talking to. Tell them the timeline, what they've said, what they look like. If you find yourself defending the relationship, or skipping the parts you know will sound bad, that's the signal. Real connections don't require a defense.
For more, see our writeup on what to look for in a dating app, or how to write a profile that signals you're a real person too. If you want to try a platform where every profile is verified at the door, join the open beta.
Founding $9.99/mo, locked forever · Available across Texas
Create Your Free Account