Romantic scammer detector: the potential of AI in dating apps to fight against fake profiles

With love in the air for the arrival of Valentine’s Day, romance is fueled on online dating apps. But what happens if instead of finding your better half you stumble upon a romantic scam? Have you ever wondered how dating apps address security challenges? In this note you learn how AI is helping to ensure a safer online dating experience.

As Kaspersky reports, a percentage of Peruvians (15%) have reported being victims of their ‘match’ sharing screenshots of their conversation without consent, threatening to expose personal information and leaking their intimate photographs.

In the midst of this panorama, dating platforms like Bumble are turning to artificial intelligence to combat online security risks.

They recently presented Deception Detector and to understand how it works, Trade spoke with Javier Tuiran, director of Latin American Communications for the app. In addition, Kaspersky gives us some recommendations to ensure a good experience in dating apps.

LOOK: Valentine’s Day: these are the applications to meet people and with which you can plan a date

How Deception Detector works

It is a tool powered by artificial intelligence designed to identify fake profiles, spam and scams. Its objective is to prevent users from encountering malicious content, so it acts preventively.

Thanks to machine learning, Bumble’s AI knows how to identify certain behaviors and factors that are related to deceptive profiles. Now, this is thanks to the data that the dating platform has collected since 2014, the year it was founded.

“We are talking about 10 years of analyzing reports of fraudulent accounts, of making reports of bad actors or bad behavior within the application. and all that Surveillance and monitoring work in recent years has allowed feeding the machine learning model”says Javier Tuiran.

Bumble applies AI for a better user experience (Top Class Actions)

Some of the indicators that ‘Deception Detector’ reviews and that helps validate the veracity of a profile are the type of words used in a conversation, which is based on reviews of chats reported as fake profiles. Also, take into account user behavior as the frequency that slides either to the right (I like it) or left (I don’t like it).

What happens after the AI ​​identifies a fake profile? According to Javier Tuiran, After detection by the tool, it is reviewed by a human moderator to finally permanently block the account in question.

“In the three months that we have been testing, we have found that Deception Detector has achieved identify up to 95% of false, misleading and reported profiles as spam”, it states.

Regarding the impact of this technology, Bumble also claims that the number of reports of fake accounts has been reduced by 45% during the months of introduction.

AI as an ally in online security

Bumble’s goal with the use of AI in its systems is ensure real and secure connections with trustworthy accounts who are looking for the same thing as the rest of the users, says the firm’s communications manager.

He explains that they are aware of the concerns people have when using their application thanks to a global survey they conducted at the end of 2023.

“The data told us that one of the main user concerns from Bumble and in general from the category of different applications It has to do with the authenticity of the connections that they were generating”.

LOOK: Cyber ​​scams can turn the search for love online into a nightmare

And at the same time recognize the advances of artificial intelligence and its uses within dating apps. For example, algorithms that help match you with more compatible people based on shared tastes (musicals, hobbies, etc.).

“So We ask ourselves why not use this same potentialthis power that comes behind AI to help people create safer connections, more authentic connections”explains Tuiran.

This isn’t the first time the dating app has leveraged AI. In 2019 they introduced the ‘Private Detector‘, which detects inappropriate or intimate photos to automatically blur them. And finally it is the user who decides to see the image or reject it and at the same time report the account.

“Without a doubt we are going to continue innovating and continue trying to bet on how we can continue using artificial intelligence to create more secure connections”declares the spokesperson.

The risks of dating app

Now, the launches of AI tools such as those reviewed in this note occur in a context of deception and fraud within dating applications. Do you know what these risks are?

Trade spoke with Eduardo Chavarro, Incident Response specialist at Kaspersky, who clarifies that in this type of platforms “we end up becoming a catalog and that catalog requires information”. That is, if one decides to create a profile to be part of the application, it is necessary to provide personal data such as telephone number, email, full names and photographs. Also, if the user decides so, it is linked to social networks.

“Even some applications enable georeferencingit basically means taking your exact coordinates and publishing them constantly“, warns the specialist.

All of these elements represent a risk where strangers can quickly find out about you. This represents an opportunity for deceptive profiles to steal money, personal data and even the identity of innocent people.

According to a global survey carried out by Kaspersky, 51% of Peruvians have used a dating app on some occasion and about 23% of them admitted to having been scammed or deceived with a fake profile.

Likewise, two out of 10 of those surveyed (19%) confessed to having been subjected to doxing, the practice of making public and uploading a person’s private information to the Internet.

Furthermore, the report adds that the three main risks that Peruvians have encountered in the dating app are: Malicious links or attachments, fake profiles, and identity theft.

LOOK: Bard vs. ChatGPT: the differences and similarities between today’s most popular AIs

Regarding the use of tools with AI, the spokesperson for the security company points out that The apps with the greatest presence among users are the ones that develop security strategies and schemes the most for the benefit of the user. “But there are applications that continue to appear and that still present some types of risks. In an analysis we did on the security of online dating apps, we found that 6 out of 9 applications have security problems”, he maintains.

Now, the expert indicates that although these models are trained with a series of concepts so that they can decide what is good and what is bad, they can also make mistakes. Thus “Human intervention continues and will continue to be a determining factor”.

Chavarro details that AI systems carry out all the exhausting work, but that finally the human staff will evaluate a smaller and more specific number of situations reported.

LOOK: Valentine’s Day: the most romantic songs on Spotify

Tips to be safe on dating apps

Although technologies are responsible for ensuring that their platforms are safe for those who use them, users also have to do the work of prevention. The Kaspersky specialist suggests the following points.

Source: Elcomercio