← Back to the Field Guide

Is My Match Real? The Four Checks That Still Work

By Cosmin · Published 26 July 2026

If you are asking the question, ask it properly rather than talking yourself out of it. Wanting a match to be genuine is not evidence that they are, and quietly looking for reassurance is how people spend three months confirming what one twenty-minute check would have settled.

The good news is that verification is quick and mostly unembarrassing. The awkward news is that the advice most people remember — reverse-image search the photos — no longer decides anything on its own. Faces can be generated on demand, which means a picture appearing nowhere else on the internet is no longer proof that it belongs to the person sending it. The checks that survived that shift are behavioural, and there are four of them.

Check 1 — Live, unscripted interaction

This is the strongest test available and the only one that is genuinely hard to defeat. A short video call, in which you ask for something that could not have been prepared in advance:

• Wave, or hold up three fingers next to your face.
• Turn the camera around and show me the room.
• Say the name of the street you grew up on.

The point is not the specific request — it is that it is live, arbitrary and physical. Recorded footage cannot comply. Real-time face-swapping exists but struggles with fast movement, a hand crossing the face, or an unexpected instruction, and it does not survive a camera being flipped to the room.

What matters at least as much as the result is the response to being asked. Asking for a quick video call before meeting a stranger from the internet is an entirely normal thing to do in 2026. Someone real finds it reasonable, or teases you about it, and does it. A refusal wrapped in hurt feelings — “I can’t believe you don’t trust me” — is doing two jobs at once: it avoids the test and makes you feel unreasonable for having set it. The technical side of what can and can’t be faked is covered in the AI dating scams guide.

Check 2 — Reverse image search, and its blind spot

Still worth two minutes, because a great many fake profiles are still built from photographs lifted from someone else’s real account. Save a photo, drop it into Google Images, TinEye or Bing Visual Search, and look for the same face attached to a different name.

The essential thing to understand is what the two possible outcomes actually mean:

• A hit is close to conclusive. The same face under another name, or on a stock library, ends the conversation.
• A miss proves almost nothing. It rules out lifted photos and nothing else. A generated face has no earlier existence to find, and a real person’s private photos won’t be indexed either.

Treat it as a filter that catches the lazy version, not a clearance. Signals worth noticing in the images themselves: a set with no ordinary photographs in it, no other people, no repeated background, and nothing casual. Small oddities in hands, ears, jewellery and text in the background are still common in generated images, though they are disappearing fast enough that their absence means nothing.

Check 3 — Cross-platform history

A real life leaves a long, boring trail. What you are looking for is not a polished presence but an unglamorous one with time depth: an account with years of ordinary posts, tagged by other people, with comments from friends who are themselves plainly real.

• Age of the account matters more than the number of followers. Followers are cheap; four years of dull posts are not.
• Other people’s content is the strong evidence. Being tagged in a friend’s photo from 2022 is much harder to manufacture than anything on their own profile.
• Consistency across platforms. Same name, same job, same city, same face, no timeline that contradicts itself.
• Verifiable specifics. A named employer whose website lists them; a university; a professional register.

The classic gap in a fabricated identity is a biography engineered to be uncheckable — working offshore, deployed abroad, contracting somewhere remote, recently widowed and travelling. Those stories are not automatically false, but they are chosen unusually often precisely because they explain away every absence at once.

Check 4 — The money tripwire

This one is not really a check. It is a rule you set in advance and do not revisit while emotionally involved, because the whole design of a romance scam script is to make the request arrive at the one moment you most want to help.

Two variants worth naming in advance so you recognise them at speed. The first is the request that is not for money at all but for a favour that moves money — accepting a transfer, forwarding a parcel, opening an account. That is money laundering, and the legal exposure lands on you. The second is the match who never asks for anything and instead teaches you to invest on a platform they recommend, where the balance goes up and the withdrawal eventually doesn’t; pig butchering is built specifically to defeat people who are confident they would never fall for a money request.

What a clean result does and doesn’t mean

Passing all four checks means you are talking to a real person who is who they say they are. It does not mean they are single, kind, honest about their intentions, or good for you. Verification and character are different questions, and the second one is answered over weeks by ordinary behaviour — plans kept, questions asked and remembered, a no accepted the first time. That is what the green flags guide is for, and it matters more in the long run than anything on this page.

Conversely, one failed check is a reason to slow down, not necessarily to conclude fraud. People do have broken cameras and private lives. The pattern is what counts: a real person eventually clears a check; a fabricated one produces a new reason every time.

If a check fails

1. Stop sending anything. No money, no further photographs, no identity documents, no address.
2. Don’t announce your suspicion. Telling them first mostly gives them a head start on deleting the account and the evidence.
3. Screenshot everything — profile, messages, photos, any payment details or account names they gave you — before you block.
4. Report the profile in the app, then block. Reporting is what gets the account removed from other people’s matches.
5. Report to the authorities if money, documents or intimate images were involved: reportfraud.ftc.gov and ic3.gov in the US, Action Fraud in England, Wales and Northern Ireland, Police Scotland on 101.

If money has already gone, the aftermath has its own playbook — what to do after you have sent money covers reporting by payment type, the realistic odds, and the follow-up scam that targets people who have just lost money. If intimate images are being used as leverage, the sextortion guide is the one to read first, and the headline is: do not pay, and it is not as hopeless as it feels tonight.

Scale, for perspective rather than alarm: the FTC recorded 55,604 romance-scam reports and about $1.16 billion in reported losses in the first nine months of 2025, with a median individual loss of $2,218. Reported cases are a fraction of real ones. This is a common, industrial crime that happens to careful people — being targeted says nothing about your judgement.

This is general safety guidance, not legal advice or a guarantee. If you are in immediate danger, contact your local emergency number.

Keep reading

• Anatomy of a romance scam — the industrial script beat by beat, and the money rules that beat every variant.
• AI dating scams: deepfakes, voice clones and chatbot matches — the verification moves that still work when reverse-image search doesn't.
• You sent money to a scammer: what to do now — the first hour, reporting by payment rail, honest recovery odds, and the second scam that follows the first.
• First-date safety without the paranoia — the seatbelt checklist: verify, venue, transport, check-in, exit scripts.