Russian Developers Train Neural Network to Flag Corruption Risks
The system automatically signals suspicious patterns, detecting both direct bribe offers and indirect warning signs.
Russian researchers have developed an AI system designed to identify corruption risks during real estate inspections. The tool was created by specialists at the Artificial Intelligence Research Center of the Institute of Social Sciences and the Institute of Law and National Security at RANEPA.
The system analyzes audio and video recordings from inspectors’ body cameras. It looks not only for explicit bribe offers but also for indirect signals, such as the handover of envelopes or the display of a payment amount on a phone screen.
Integrating Algorithms Into the Capital’s IT Framework
Researchers have prepared a detailed methodological framework that will serve as the basis for updating regulatory documents and inspection guidelines.
The successfully tested algorithms are expected to be integrated into the existing digital systems of the Moscow Department of Information Technology. In the future, the tool could be scaled across all city inspections.
Officials expect the AI instrument to increase transparency across government oversight bodies and reduce the impact of the human factor.
