Strengthening Digital Security: Sber Patents Deepfake Detection Technology

A Growing Market for Media Security Solutions

Participation in the 'Dialogue on Fakes 3.0' forum, included in UNESCO’s Global MIL Week calendar, underscores the scale and importance of combating synthetic content. The event serves as a key venue for discussing media literacy and fact-checking, expanding awareness among policymakers, tech professionals, and the general public.

For Russia’s IT industry, this signals a strengthening of competencies in AI Trust & Safety, antifraud systems, and media security — forming a promising new market for telecom, media, banking, and fintech sectors. Sber’s proprietary patents and technologies will play a central role in that development. Citizens will benefit from enhanced protection against financial and reputational harm caused by video and voice forgeries.

At a national level, these innovations reinforce Russia’s technological sovereignty in content protection and encourage future legislation.

Globally, Sber’s deepfake detection technologies could be exported to international Trust & Safety markets, supported by the company’s engagement in UNESCO’s initiatives.

“Our patented inventions allow us to effectively identify information and cyberattacks that leverage deepfake technologies. Synthetic videos used in attacks against Sber and our clients will be detected at the earliest stages.”

Stanislav Kuznetsov

Deputy Chairman of Sberbank

Rising Demand for Antifraud Technologies

Sber envisions its deepfake detection solutions as integral to critical infrastructures — from telecommunications and broadcasting to social media, messaging platforms, and even mobile devices. Implementation will require appropriate legal frameworks, which Russian lawmakers are already developing.

Since 2022, Sber has held patents in this field and is preparing to export its technologies via licensing and SDK models. A draft federal law is now being discussed in the Russian parliament to regulate or prohibit malicious deepfakes, including mandatory content labeling and detection. Given the confirmed surge in AI-based fraud attempts targeting banks and e-commerce platforms, demand for antifraud and verification tools from financial institutions and government agencies is projected to grow rapidly.

Cyberattack Surge and Defensive Technologies

In 2022, Sber patented its deepfake detection technologies, achieving up to 98% accuracy in independent testing. The system is designed to counter cyberattacks that bypass facial recognition and liveness detection systems, as well as to detect information attacks, fake news, and threats to video conferencing security.

A key feature is its ability to analyze footage containing multiple faces, identifying synthetic identities and neutralizing attempts to evade deepfake detectors. By 2024, cyberattacks on Russian organizations had risen dramatically: attacks on cloud and hybrid infrastructures tripled from 29 million in 2023 to 84 million in 2024. Overall, cyber incidents increased by 2.5 times, with a 30% rise in employee device-based intrusions. Government agencies were the primary targets, followed by the transport sector. Attackers are increasingly using coordinated tactics — one group provides access, while another delivers the payload.

In 2025, Russia’s Federation Council began working on legislation to define and regulate deepfakes, led by Speaker Valentina Matvienko. The initiative will establish legal definitions, distinguish creative use from malicious manipulation, and propose mandatory deepfake labeling. By October 2025, Sber’s Deputy Chairman Stanislav Kuznetsov reported an 'explosive rise' in deepfakes, noting that generating a convincing fake now takes only seconds. He emphasized the need for rapid integration of detection tools across consumer and enterprise systems.

AI Content Labeling and Export Potential

Sber’s announcement confirms the maturity of domestic auto-detection technologies and their readiness for large-scale deployment. The topic has shifted from research to industrial implementation, driving regulatory and operational change.

Within the next one to two years, Russian telecom operators, OTT platforms, banks, and media organizations are expected to launch pilot projects. Mandatory detection and labeling of AI-generated content will likely become part of platform compliance requirements. The financial sector, in particular, is preparing to expand procurement of antifraud solutions as deepfake-related scams escalate.

Over the next three to five years, Russia’s Trust & Safety solutions — including SDKs and APIs for video and audio analysis — could gain a foothold in export markets with active content regulation policies. Existing patents, scientific foundations, and public readiness to share technologies provide a solid basis for this expansion.