Fraudsters generate other people's voices using text-to-speech and

[tanjimajuha20]

Specialists from the AIRI Institute of Artificial Intelligence and the Moscow Technical University of Communications and Informatics (MTUCI) have developed a system that uses AI to identify signs of synthetic voice generation used by telephone scammers. This was reported to TASS by the AIRI press service.

"[Modern technologies] make georgia whatsapp number databa se it possible to create synthetic voices that are already very difficult to distinguish from real ones. The advantage <…> [of AI models such as the presented solution] is their ability to take into account the context and knowledge of voice data, allowing them to more effectively distinguish between a genuine voice and a fake one. These networks not only recognize fakes with high accuracy, but are also able to adapt to new types of threats. The introduction of such advanced methods significantly increases the level of security and protection against attacks based on voice substitution," the institute's press service quotes Oleg Rogov, head of the AIRI Trusted and Secure Intelligent Systems research group, as saying.

voice processing technologies. The authors proposed a solution for analyzing audio sequences, the performance of which is twice as high as the model created earlier by scientists in South Korea and France. The Russian development can be used to combat telephone fraud or improve the security of systems using voice authentication, the scientists believe.

"Integration <…> can be carried out in various ways: from the implementation of a separate software solution to integration into existing security systems. <…> The need for such technologies is high, given the growing threat of attacks using synthetic voices," said Grach Mkrtchyan, head of the MTUCI Intelligent Solutions research department, as quoted by the institute's press service.

The AI ​​model was among the top ten solutions in the international ASVspoof 2024 Challenge.