Tomsk Polytechnic University Trains AI to Assess the Quality of Industrial Safety Briefings
The system operates with more than 90% accuracy.
Researchers at the Industrial AI Laboratory at Tomsk Polytechnic University (TPU) are developing an intelligent system designed to improve safety at industrial facilities. The system assesses the quality of employee safety briefings and uses an algorithm to distinguish between speakers' statements. This makes it possible to determine whether the person responsible for delivering a briefing has covered all the required safety topics.
The system has been tested on 1,000 audio recordings. Researchers evaluated the performance of its algorithms across the entire workflow, from uploading audio and video files, recognizing speech, and identifying individual speakers to analyzing briefing content and generating summary reports. Accuracy exceeded 90% across all key areas.
