Russian Students Develop AI to Classify Citizens’ Requests
The tool is designed for use in the public sector to improve communication and analytics.
Students at St. Petersburg State University of Aerospace Instrumentation have created a machine learning-based program to process citizens’ requests. The system can be used in public administration to automatically determine which agency should receive a given request. Regional officials say there is strong demand for higher-quality analytics of incoming messages.
The students tested neural network models capable of text recognition. The system was trained on anonymized archival data. It works as a “sorter”: a file containing messages is uploaded, and the AI assigns each one to a relevant sector.
Tested in Practice
The solution was piloted at the Center for the Development of Civil Society Institutions. The neural network was trained on 140,000 records and can already accurately route messages related to healthcare, transport, roads, education, or urban improvement. The system is set to be integrated into other government institutions in St. Petersburg. Officials expect the time required for preprocessing messages to fall, enabling more thorough analysis and improving feedback to residents.
