Russia Develops Chatbot for Detecting Fires
The neural network analyzes photos to identify where and why a fire may have started.
Russia has developed a neural-network-based tool for firefighters, the Emergencies Ministry said. The smart chatbot analyzes photos and can determine the seat of a fire as well as its preliminary cause. The system was developed jointly with the nonprofit Institute of Artificial Intelligence.
According to TASS, the system, called OmniFusionFires, is capable of identifying the initial ignition point. To do this, an Emergencies Ministry official uploads photos from the scene into the chatbot.
For Official Documentation
The information generated by the system can then be used for further investigation and for preparing official expert conclusions.
In 2025, the Emergencies Ministry, together with Yandex, also launched a neural network designed to describe emergency scenes. Such descriptions are required for official reports and documentation, a task that can take a human an entire day. The neural network, by contrast, scans photos from the scene and produces a text report within minutes. The integration of neural networks into the ministry’s work began in 2024.
