Spotting fake dating profiles

How a fake profile differs from a bot

These are two different problems and should not be confused. A bot is an automated account that sends out identical messages and usually steers you to an external site. It gives itself away within a couple of exchanges: it answers beside the point, ignores your question, and drives on with its own script.

A fake profile is a living person behind someone else’s photograph. They answer sensibly, remember details of your life, and keep a conversation going for months. That is the dangerous one, because bots ask you to click a link while this one will ask for money.

The practical consequence: tests like “ask something odd, a bot will not answer” work only against the first category. Against the second you need a check on identity, not on wit.

How to check a photograph in two minutes

A reverse image search is the most effective of the free methods. You save the photograph from the profile and upload it to Google Images or TinEye. If the shot turns up in accounts under other names, on a stock site, or in a five-year-old article, the conversation is over.

It is worth checking two or three photographs rather than one. Fraudsters often take shots of one person from a real account: one photograph may turn up nowhere, while the second leads to the original.

Generated images are a story of their own. Hands and ears used to give them away; they no longer do. But a generated photograph turns up nowhere else — and that in itself is a signal: a real person active on the internet usually leaves at least some trace.

Nine signs that recur most often

Individually none of them proves anything. A person really can have a bad camera, and really can work abroad. What matters is how many coincide: three or more is a reason to stop.

• No mutual acquaintances, and they wrote first
• The conversation moves quickly to WhatsApp or Telegram
• A declaration of feelings before you have met
• An occupation from a short list: doctor, engineer, soldier, sailor — always abroad
• A video call never works out; the camera has been broken for months
• The photographs look too professional for an ordinary person
• Turns of phrase that look like machine translation slip into the writing
• They ask you not to tell those close to you about the relationship
• Sooner or later an urgent money problem appears

Why the occupation in the story is always the same

The Finnish police, in their piece on romance fraud, list the type almost word for word: a doctor on a mission, an engineer on a rig, a soldier at a base, a sailor on a voyage. The list is short and has not changed in years.

The reason is utilitarian. Such an occupation answers three awkward questions at once: why you cannot meet, why the connection is bad, and why someone on a good salary suddenly has no access to their money. One story plugs every hole in the scheme.

Checking it is simple. Ask about the place they are supposedly in: today’s weather, the district, the time. The story is usually worked out at the large scale — job title, company, history — and falls apart on the everyday.

What the services themselves verify, and what they do not

The spread is enormous, and services do not like to talk about it plainly. Some platforms check only that an email address is unique. Some ask for a phone confirmation. Few require a selfie in a set pose and compare it against the profile photographs.

Selfie verification is the only common measure that genuinely prevents the use of someone else’s photographs. Everything else can be worked around in five minutes.

What to look at is not the word “verification” on the front page but what exactly is being verified. The phrase “all profiles are checked by a moderator” means nothing: the moderator looks at whether the photograph contains prohibited content, not at whether it is the right person.

What to do about AI photographs and deepfakes on a video call

Over the past two years checking has got harder, and the old advice works less well. A generated image is no longer given away by an extra finger or a crooked ear — those artefacts were cleaned up long ago.

But an inverse sign has appeared, and it is even more reliable. A real person who lives actively on the internet leaves a trace: a photograph surfaces somewhere, at least once. A generated picture turns up nowhere, ever. An empty reverse-search result used to mean “all fine” and now means “unclear” — and calls for a second check.

A second practical technique: compare two or three shots with each other. A generator is bad at keeping small details consistent — a mole, the shape of an earring, the pattern on a mug, the number of steps in the background. A person stays themselves across different photographs; a picture does not always.

The same goes for a video call, only there the check is active. A real-time deepfake still copes badly with three things: a sharp turn of the head into full profile, passing a palm in front of the face, and bringing an object up to the face — a mug, a phone, a sheet of paper. At those moments the mask usually tears or jumps.

You can ask for this in passing and without explaining. Someone with nothing to hide will turn their head and not think about it. Someone with something to hide will find a reason, or suddenly lose the connection.

How a fake profile gives itself away specifically in Finland

There are signs that do not appear on international lists because they are tied to the local context. On Finnish platforms they work better than the general ones.

The first is geography in small details. The profile says Helsinki, but the person does not know the difference between Kallio and Punavuori, does not understand what an Alepa is, and has no opinion about rents. For anyone who actually lives here, that is background you talk about without thinking.

The second is a combination that does not add up: flawless Finnish in the messages together with a story of “working abroad for a year now”. Finnish at that level usually means the person grew up here, and such a person is rarely unreachable for a meeting.

The third is time. Someone supposedly in Tampere answers at night and is silent during the day. A time zone difference shows in the rhythm of the messages better than in any words.

The fourth is money and institutions. A real Finnish story about financial trouble runs into specifics: Kela, ulosotto (enforcement), a bank. A made-up one usually gets by on generalities about a “frozen account”, because the author does not know the details.

Is a fake profile a crime in Finland?

In itself, as a rule no, and it is worth understanding that in advance so as not to build up expectations of what a report will achieve.

But there is a separate offence. Since 4 September 2015 the Finnish Criminal Code has contained a provision on identity theft — identiteettivarkaus, chapter 38 section 9a. It applies where a person unlawfully uses another’s personal or identifying data in order to mislead someone and cause the injured party financial or other damage that is more than minor.

Two details make it weaker than it looks. The penalty is a fine only; the provision carries no imprisonment. And it is a complainant offence: the prosecutor brings it only at the injured party’s request. That is, the person whose photographs were stolen has to report it themselves.

Once money is involved a different provision applies — fraud — and it is more serious. So a report should describe not only the fake profile but the whole financial part, if there was one.

The practical conclusion is simple. If your photographs are being used in someone else’s profile, the report should come from you rather than from whoever found the profile. And if you were harmed as the person on the other end of the conversation, what supports your case is not the fake profile itself but what followed from it.

What to do if the profile turns out to be fake

First, do not tell the person. The moment a fraudster realises they have been found out, they delete the account along with the messages. Screenshots first, everything else afterwards.

Second, report the profile inside the service. This is not a formality: platforms do delete such accounts, and the next person will not see it.

Third, if money was transferred the order is different and urgent: the bank first, then the police. If there was no money, a report is still worth filing, because cases are built out of episodes.

If you feel you have become too suspicious

A side effect that articles about fraudsters do not write about, although almost everyone who reads such articles runs into it.

Having read a list of signs, a person starts seeing them everywhere. They did not take a video call — suspicious. They work abroad — suspicious. They write well — also suspicious. In the end the checking eats up the pleasure of meeting someone, and sometimes the meeting itself.

What helps here is going back to how the signs work. They only work in combination: three or more together. Individually each is explained by ordinary life, and that is not a softening caveat but a direct consequence of how these lists were compiled.

The second thing that helps is separating checking from talking in time. The check takes five minutes: a reverse image search and a short unannounced video call. Once done, the question is closed and you simply talk. Constant background checking of every word adds nothing to the result of those five minutes.

Third: suspicion is no substitute for a boundary. Even a completely real person can turn out to be unpleasant, and what protects you from that is not a profile check but a readiness to end the contact at any moment without explaining.

And the thing worth saying plainly. Most people on dating platforms are ordinary people, and the statistics bear that out: on police figures for 2024 there were 460 reports of romance and document fraud in the whole country. That is a lot for those affected and little relative to the number of people meeting. Caution makes sense as a procedure, not as a state of mind.

What not to do

Do not run background checks on a person through databases or order an “identity check” from dubious services. Such services usually sell you a fifty-euro report compiled from open sources, and not infrequently turn out to be frauds themselves.

Do not try to expose the person through the conversation. It achieves nothing: they do this professionally and have done for years, while you lose the time in which evidence could have been preserved.

And do not send “a little, just to test”. Testing by sending money does not work — a small first transfer is part of the scheme, and it is exactly what shows the fraudster that you are willing to pay.

Ways of checking: what works and what does not

The signs and the scheme follow poliisi.fi material, updated 30.5.2024.

Frequently asked

Sources

• Poliisi: Rakkaushuijaus — tästä se yleensä lähtee — the type of story, the phases of the scheme
• Traficom: Näin suojaudut nettihuijaukselta
• Poliisi: filing a report
• Eastern Finland police: statistics 2025
• RIKU: what identiteettivarkaus is and when it is punishable — RL 38:9a, a complainant offence
• Ministry of Justice: identity theft as an offence in its own right — in force since 4.9.2015, penalty a fine
• Yle: 53 per cent of Finns have tried a dating service