Neural networks of love: how scammers use AI on dating sites
Scammers can use artificial intelligence on dating sites to deceive Russians, experts have warned about this. Neural networks allow attackers to automate what they used to do manually, while maintaining the illusion of human contact. For more information about how scammers use AI on dating sites, how dangerous such schemes are and how to protect yourself from them, read the Izvestia article.
Why is the topic of dating sites interesting to scammers
Dating sites attract intruders with a high level of trust between participants who are emotionally involved and open to communication — this is what allows scammers to build long-term manipulative scenarios, including extorting money under the guise of a romantic relationship, Andrei Goryachev, director of public relations at the Security Code information security company, says in an interview with Izvestia.
— In addition, there is a lot of personal data on such sites — they can be used for further attacks or blackmail, — the expert notes.
A person comes to the dating platform with the expectation of finding a partner, he is ready to share personal information, photos, and talk about his life, which reduces the criticality of perception and makes people vulnerable to manipulation, adds Irina Mezheneva, a leading analytical engineer at the laboratory of artificial intelligence technologies at the analytical center for cybersecurity at Gazinformservice.
It is also important that many users have no experience in detecting fraudulent schemes in the digital environment, especially among the older age groups who actively use such services. Previously, the tricks of scammers required time and manual work — correspondence, maintaining a legend, creating a plausible story. AI automates this process and allows you to conduct hundreds of conversations simultaneously at minimal cost, the analyst emphasizes.
How scammers can use AI on dating sites
Scammers on dating sites often use neural networks to create an image of an ideal partner — with a suitable appearance, voice and manner of communication, MTS Defender experts say. The attackers' goal is to build a long—term trusting relationship with a potential victim in order to gain access to her money or personal data.
"First of all, neural networks generate realistic images of non—existent people for fake profiles," says Yakov Filevsky, an expert in sociotechnical testing at Angara Security, in an interview with Izvestia. — There is no need to steal someone's photos anymore: an attractive avatar is created in seconds and is not found by reverse search. At the same time, the correspondence itself is automated.
According to the source, chatbots based on large language models are able to communicate with hundreds of people simultaneously, showing empathy and adapting to the writing style of potential victims. All this gradually creates a sense of trust on the part of the latter, said Leonid Bezvershenko, senior expert at Kaspersky GReAT. At the same time, there are many romantic scam schemes where money or data is extorted from the user under various pretexts.
For example, a scammer can talk about his successful investments (often in cryptocurrency) and offer to "earn together." To lull vigilance, he can help to withdraw a small amount, but then persuade them to invest all their savings, promising huge profits that the victim will never see, the specialist explains. Automated campaigns can also be implemented, where part of the communication is conducted using AI, and the "connection" of a real person occurs already at the stage of requesting money.
— A separate scenario is audio and video clips to simulate a "live" dialogue, — says Leonid Bezvershenko. — At the same time, the very idea of "identity forgery" is not new: before AI, attackers also used other people's or edited photos, posing as other people. AI has only simplified this process, allowing fakes to be made more plausible and widespread.
What AI techniques have you encountered on dating sites?
Through images of girls created using neural networks, fraudsters had previously lured users into deception schemes involving false investments, installing malware on phones to steal money from bank accounts under the guise of dating apps, sharing remote movie viewing, and also under the pretext of fake dates, a senior analyst told Izvestia. Department of Digital Risk Protection of F6 company Maria Sinitsyna.
—Only three scam groups that use the FakeDate scheme stole more than 330 million rubles from Russian users from January to October 2025," the expert notes.
In practice, there have been cases when a character completely generated by a neural network conducted parallel affairs with dozens of people and collected money for "treatment" or "for a meeting," adds Andrey Goryachev. In one of the cases, the scammers used AI to synthesize voice and short video messages at the victim's request, which convinced her of the reality of the interlocutor. Attacks were also recorded, where the neural network itself adapted the legend to the victim's biography, extracted from her social networks.
According to Yakov Filevsky, the target audience of such scenarios is most often people in a period of social or emotional vulnerability: after a divorce, loss, relocation, or prolonged loneliness. Age or level of knowledge takes a back seat here. In real practice, financially literate and successful adults are easily manipulated when the algorithm hits their actual psychological need.
"The consequences of such attacks are twofold: on the one hand, material damage is caused — attackers adapt to bank fraud, so they try to split up the amounts, stretching the theft into a series of small payments," warns the source. — On the other hand, the victim of deception receives severe psychological trauma. Due to feelings of guilt and shame, seeking help is delayed, which leads to the loss of larger amounts.
How to protect yourself from AI fraud schemes on dating sites
In order to protect themselves from neural network schemes on dating sites, experts interviewed by Izvestia advise following certain safety rules. In particular, Maria Sinitsyna urges to monitor the appearance of new fraudulent schemes in order to minimize the chance of falling for them.
— If you are being sent recorded videos when communicating with people you have never seen in person and only know online, then ask them to call you through a video communication app to talk in real time, — the expert recommends. — Refusal in this case is a sign of possible deception.
She also adds that AI allows attackers to use "masks" even during a video call, so to verify the authenticity of the interlocutor's image, ask him to do an unexpected action: for example, quickly swipe his hand in front of his face to the right or left. If the picture starts to "crumble", there will be image distortions in the area of the interlocutor's face — this is a sign of a possible deepfake.
Given that AI tools for creating deepfakes are constantly being improved, these verification methods do not allow us to guarantee one hundred percent recognition: it is always worth paying attention to the context of communication, the expert notes. If the interlocutor forces you to take any actions related to the disclosure of your data or financial transactions, this is a likely sign of fraud. In the early stages, it is better to keep communication inside the application and be wary of attempts to transfer the dialogue to third-party channels.
"Suspicious persistence, a quick transition to financial topics, or requests for help are a reason to be wary,— concludes Leonid Bezvershenko. — In general, if in doubt, it is better to stop communicating and use the built-in complaint tools. Additionally, you should install a security solution with the function of blocking phishing links on your device.
Переведено сервисом «Яндекс Переводчик»
