AI Blocks Suspicious Calls Before Users Pick Up
Systems built on artificial intelligence can now block suspicious calls before a user answers. According to Yandex experts, algorithms evaluate more than 300 behavioral signals, including call frequency, duration and volume, as well as indicators of SIM-box activity. Decisions are made within seconds and rely on anonymized data.
Russia’s fight against phone and online fraud has entered a systemic phase. In 2025, the Central Bank of Russia blocked more than 69,000 phone numbers and nearly 40,000 fraudulent websites and pages. Starting in March 2026, the state system Antifrod (Anti-Fraud system) will launch to connect market participants and enable the exchange of signals about suspicious activity.
For the IT sector, this marks a shift from basic spam filtering to predictive analytics. Algorithms now detect behavioral anomalies and signs of number spoofing. That is increasing demand for machine learning engineers, antifraud platforms, telecom analytics and technologies for processing anonymized data.
For users, this translates into protection before the first ring. The system can block suspicious calls or issue warnings during conversations, reducing exposure to social engineering attacks. At the national level, the architecture reduces pressure on banks and telecom operators while making the digital environment more resilient, aligning with a broader global shift toward AI-driven fraud prevention.
Digital Hygiene Infrastructure
In Russia, antifraud solutions are rapidly evolving from optional services into a baseline security standard. The market is already showing measurable impact: in 2025, MTS blocked more than 3 billion fraudulent calls, attributing the result to AI deployment and tighter regulation. These tools are becoming part of core digital hygiene infrastructure.
Further development is expected across several directions, including a move from number verification to real-time conversation analysis, integration with the Antifrod system, stronger detection of SIM-box operations and the emergence of family and insurance services built around antifraud capabilities. This shift turns fraud protection from a supporting feature into a standalone market product.
Export potential remains limited, largely because these systems depend on national regulations and infrastructure. However, behavioral analytics and fraud pattern detection technologies are broadly applicable. Russian companies can offer antifraud platforms and B2B modules in CIS and partner markets where demand for such systems is growing.
Global Shift
In 2024, the government announced plans to create a unified platform to combat phone fraud as part of digital security efforts. In 2025, the initiative moved into a regulatory phase, culminating in the launch of the Antifrod system in March 2026. This indicates that the current developments are part of a longer-term effort to build a national antifraud infrastructure.
Meanwhile, the market has moved from pilot projects to industrial scale. In 2025, MTS reported blocking 1.5 billion fraudulent calls in six months, and by early 2026, that figure had reached 3.17 billion for the full year. This shows the sector has moved beyond experimentation to large-scale filtering of massive traffic volumes.
In 2025, MegaFon developed the voice assistant Eva to block spam and fraudulent calls while also serving as a family protection tool. Countering phone fraud has become a shared strategic priority among major telecom operators.
In March 2025, Google introduced the Scam Detection AI feature for calls and messages on Android. On Pixel devices, the technology runs locally, while on Pixel 9 and newer models it uses the Gemini Nano module. This highlights a global trend: fraud protection is shifting from telecom operators to end-user devices. U.S.-based Hiya is following a similar path, advancing real-time fraud protection, fake voice detection and AI-powered call assistants. The market is moving beyond simple number filtering toward a broader class of voice security services.
Building a Multilayered System
The development of antifraud systems reflects the maturity of the domestic market. Protection against phone fraud has become a unified, multilayered system combining AI tools used by telecom operators and banks with a state-level coordination layer. At the same time, losses from fraud reached 29.3 billion rubles (approximately $320 million) in 2025, signaling a new phase of competition between attackers and defense systems. Each technological breakthrough on one side triggers adaptation on the other, making antifraud an ongoing process rather than a fixed endpoint.
In the near term, predictive call scoring, real-time conversation analysis and integration of antifraud signals across banks, platforms and government agencies are expected to expand. This increases the practical role of AI in cybersecurity and strengthens digital resilience, while for users it means more reliable protection even before a conversation begins.
