The promise of Discover the face of your soulmate. It fails before any discussion of algorithm quality. It fails at the source: There is no input data.. A prediction requires information about what you want to predict, but about a person you haven't met yet, there is no information at all.
This text explains the reasoning step by step, shows why the result seems to be correct nonetheless, and discusses what the idea of a soulmate does to those who believe in it—which is the most interesting part of the subject.
The problem isn't the algorithm, it's the lack of data.
Any forecasting system works the same way: it receives data about what has already happened and projects what is likely to happen. A weather forecast is based on current measurements of pressure, temperature, and wind. A movie recommendation is based on your viewing history.
Now consider what a soulmate app has to offer: a photo of you, perhaps your birthdate, and a few answers. None of that contains information about a specific person you haven't met yet—because that information doesn't exist anywhere, not in the app, not in any database in the world.
It's not a matter of waiting for better technology. A forecasting system without data about the target isn't predicting badly: it's not predicting at all.
And the meeting hasn't even been scheduled yet.
There's a second layer, and it's more fundamental. Who you'll meet depends on a huge chain of chance: the city you'll live in, the job you'll accept, the friend who will introduce you to someone, the time you took that bus.
None of this is determined the moment you open the app. There is no information to be read about an event that has not yet been defined — which would put the promise beyond the reach of any system, no matter how advanced.
Why the result seems to be correct.
There is a well-studied phenomenon in psychology known as Barnum effect. Vague and generic descriptions are perceived as surprisingly personal by almost everyone, especially when presented as the result of some method.
It's the same mechanism as generic horoscopes. "You have a need to be liked, but you're also critical of yourself" sounds specific and applies to practically anyone.
It works the same way with faces. A portrait with average features bears a resemblance to many people. After seeing it, when meeting someone, memory highlights what matches and discards what doesn't—a classic confirmation bias. The portrait didn't predict; it provided a mold flexible enough to fit several people.
The cost of believing: what the idea of soulmates provokes.
This is the part that rarely appears in texts on the subject, and it is the most useful.
The belief that there is a predestined partner tends to produce a specific interpretation of conflicts: if it's difficult, it's because it's not the right person. Those who start from this premise usually interpret disagreements as a sign of fundamental incompatibility, and not as something to be resolved.
The alternative interpretation—that relationships are built, and that difficulty is a predictable part of the process—leads to different behavior when faced with the same conflict: talking, adjusting, negotiating. It's not a matter of romanticism versus pragmatism; it's that the two beliefs produce opposite reactions precisely when the reaction matters.
There is also a practical effect: expecting to recognize the right person by an external sign — a face that matches a portrait — diverts attention from what is actually observed in someone over time.
What is really known about attraction?
Several factors consistently emerge when studying couple formation, and none of them are mysterious:
- Proximity. People form bonds with those they interact with — work, school, neighborhood, circle of friends. It's the most mundane and the most decisive factor.
- Familiarity. Repeated exposure increases likeability. Getting to know someone gradually often makes more of a difference than a first impression.
- Similarity. Similar values, lifestyle, and vision for the future make living together easier—more so than shared tastes.
- Reciprocity. Knowing that the other person is showing interest increases one's own interest.
- Moment. Two people at compatible stages of life have a much greater chance than two people at different stages.
Note that none of these factors are predictable from a photo of you. They all depend on where you are and who you meet.
What to do with this in practice
If the intention is to meet someone, what increases the chance is increasing the number and quality of possible encounters. Dating apps do this well—not because they predict, but because they expand the circle.
happn: dating app
AndroidIt's worth applying the same precautions as always: honest profile, uncovered face photo, conversation before the meeting, first meeting in a public place, and someone knowing your location. This changes the outcome more than any photo generated by an app.
SEE MORE:
- How the soulmate diagram is generated
- Compatibility in love: what can be measured?
- Safe apps for chatting with people nearby.
What do you give in exchange for the game?
It's worth looking at the transaction from another perspective. The app returns an image with no informational value and receives a set of data with real market value: full name, date of birth, city, email, and often a photo of your face. This combination is exactly what is used to create advertising profiles and, in case of a data breach, to attempt account recovery on other services.
The photo deserves separate attention. Under the General Data Protection Law, biometric data is classified as sensitive personal data, with stricter requirements for consent and purpose. An entertainment app that receives facial images needs to clearly state how long it stores them, whether it uses them to train its own system, and how you can request their deletion.
Also read the section of the terms that deals with the license to use the submitted content. It is common to find a broad assignment, allowing the company to use the submitted images for advertising and service improvement. It's not illegal, but it's the type of clause that no one would knowingly accept in exchange for a randomly generated drawing.
How to rate an entertainment app before installing it.
- Consistent permissionsGenerating an image doesn't require contacts, address book, SMS, phone number, or precise location.
- Weight and functionalityIf the app is small and everything is processed on a server, your data leaves the device each time you use it.
- Accessible privacy policy, in Portuguese or another readable language, including contact information for the data protection officer.
- Option to delete account within the app, now required by the two major stores.
- Visible billing model Before we begin: don't just find out the price after the result is blurred.
- Number of adsFull-screen ads between each tap indicate an app designed for printing only.
A simple detail separates honest entertainment from the rest: apps that present themselves as a joke usually say so in their description. Anyone promising revelation, destiny, or predictions is making a claim they can't back up, and that alone speaks volumes about the quality of the rest.
What this type of app doesn't do, bluntly speaking.
The list of limitations is short and should be stated directly, because the advertising for these services insists otherwise:
- It doesn't predict who you'll meet, where, or when.
- It does not consult databases of real people, because such a database does not exist and accessing photos of third parties for this purpose would be illegal.
- It doesn't analyze compatibility based on your face: facial features don't carry any information about romantic affinity.
- It doesn't improve with more information from you; additional data only refines the text accompanying the image.
- It doesn't repeat the same result if you retake the test, which in itself shows that there's no prediction involved.
None of this means the app needs to be uninstalled in anger. It just means it belongs to the same category as a fun camera filter, and should cost the same as a fun filter: nothing, and no data beyond what's necessary.
Step-by-step guide to playing without putting yourself at risk.
- 1. Install only from the official store and check the developer.
- 2. Use a secondary email address and avoid logging in with your main social media account.
- 3. When requesting a photo, please choose an image that you don't use on any public profile.
- 4. Decline notifications and location services on first run.
- 5. Do not register your card to see results, even in a trial period of just a few days.
- 6. When finished, delete the account within the app and then uninstall it.
The most common mistakes always appear in the same order: logging in with the main account for convenience, accepting notifications without thinking, accepting the free trial to see the image in high resolution, and publishing the result with one's own name and date of birth visible in the generated artwork.
Frequently asked questions about soulmate apps
Se eu não enviar foto, o app fica seguro?
It's better, but that's not the only issue. Without a photo, you still need your name, date of birth, email, and device identifiers. The risk is significantly reduced, and the experience doesn't change, since the generated image doesn't depend on your face to be produced.
Dá para apagar meus dados depois?
Yes, and it's a right guaranteed by Brazilian law. The request is usually within the app, in the account area, and can also be sent to the contact indicated in the privacy policy. Keep the order confirmation, with the date.
Esses aplicativos funcionam offline?
Rarely. Image generation requires heavy processing and happens on a server, which means a mandatory connection and data transfer. If an app promises instant offline results, it's simply randomly selecting from images already included in the package.
