AI vs Cyberattacks: How Neural Networks Fight Fraudsters

On October 30, 2026, the publication NGS.ru published an article about how neural networks help combat fraudsters on smartphones and computers. The note lists state protection mechanisms: self-prohibition on loans, self-prohibition on obtaining SIM cards, and the impossibility of receiving a code from Gosuslugi during a phone call. These measures operate at the legislative level and do not require the use of AI.

What distinguishes this material from previous publications on this topic is its emphasis on neural networks, however, the text contains no specific examples of AI working in the fight against cyber fraud. Instead, already known administrative tools are described in detail. For example, since March 1, 2025, a person can set a self-prohibition in their credit history on concluding consumer credit or loan agreements. It applies to banks and microfinance organizations, and if an attempt is made to arrange a loan while the prohibition is in effect, the lender must refuse. The self-prohibition can be configured under various conditions, for example, prohibiting remote arrangement or setting a full prohibition. It is valid indefinitely until the person removes it themselves. Whether a self-prohibition is set can be checked on the Gosuslugi portal.

In practice, this means that a user can independently limit fraudsters' opportunities without relying on algorithms. The self-prohibition on issuing SIM cards does not allow telecom operators to conclude a contract in your name without your knowledge. The prohibition on receiving a code from Gosuslugi during a phone call blocks one of the popular deception scenarios. These measures do not require technical knowledge and are available through government services. However, the article does not specify exactly how neural networks participate in protecting devices — for example, in call filtering or analyzing application behavior.

It remains unknown which specific AI solutions are used to combat fraudsters and how widespread they are. There is also no data on how neural networks interact with the listed government mechanisms.