Russian Companies Embrace Deepfake Detection Technology Amid Growing Concerns

Russian companies have begun to implement systems for detecting deepfakes. These technologies are designed to identify counterfeit videos, audio, and images created by neural networks. The systems operate in real-time and are currently employed in approximately 30 projects, with around half of them in the banking sector.

The detection systems analyze not only the visible signs of forgery but also assess how exactly the fake content was generated. Igor Bederov, head of the company «Internet Investigation,» explained that detection neural networks can identify which model was used in the creation process. Frequently, the responses include names of popular audio and video generators.

Fakes are becoming increasingly sophisticated, and even readily available tools can produce highly realistic forgeries, making detection more challenging. Even forensic experts may struggle to spot the differences.

Some experts suggest that these detection systems should be mandatory. For instance, «Sber» has called for a standard for deepfake recognition. However, Alexey Muntyan from the privacy specialist community believes there is currently no urgent need for this. Nonetheless, companies are preparing for an anticipated rise in counterfeit content. According to him, high-quality deepfakes in real-time demand considerable resources, but as technology becomes cheaper, the number of fakes may increase.

Authorities are also tightening regulations. A new legislative package called «Anti-fraud-2» is under discussion, which introduces additional requirements for protection against fraudsters. Communications operators and banks may face increased liability if clients suffer losses.

The Ministry of Digital Development of Russia is proposing stricter penalties for the use of deepfakes. The agency wants to classify this as an aggravating circumstance in cybercrimes. This initiative is currently in the discussion phase.