Russian developers have introduced an artificial intelligence system capable of instantly identifying scam messages. The story, cited as a report from representatives of Kuban State Technological University, highlights this development as a notable advancement in online security.
The publication explains that the program is written in Python. According to Oleg Leinweber, a developer involved in the neural network, the AI captures a screenshot of the active window every 15 seconds and analyzes the textual content in the image to detect spam or fraudulent language.
When the system recognizes a malicious message, it immediately alerts the device owner. The on-screen notification not only displays the suspicious message but also classifies it by category, depending on the type of fraud detected.
Journalists note that the new AI is compatible with computers running Linux and Microsoft Windows, while support for other platforms has not yet been introduced.
Tests indicate that the neural network was trained on a broad set of common fraud scenarios. In particular, it targets typical attempts to extract bank card data and offers related to working on online marketplaces.
The algorithm is expected to be beneficial for both individual and corporate users. Additionally, the developers envision specialized versions tailored for educational institutions and government agencies.
“Users may trial the beta version of the software this year, with a full version anticipated in 2026,” the report states, outlining a clear roadmap for deployment and expansion.
In commentary on the project, Ideco’s director Dmitry Khomutov described the success factors for the initiative. He emphasized that the neural network’s reliability and efficiency will largely determine its appeal, and that ease of integration into existing workflows is a critical consideration for organizations evaluating adoption.
Some industry experts, however, have urged further optimization to minimize resource consumption on user devices. Despite this, the overall assessment remains positive, with widespread interest in how such a tool might curb fraud and protect digital assets.
Previously, Rostelecom reported indicators about the scale of cyberattacks targeting the State Services portal in 2023, underscoring the broader context in which new defensive technologies are being developed and tested.