Web Article
GS to offer self-developed AI safety agent free of charge to SMEs
Created on April 24, 2026

GS Group has announced that it will provide its internally developed artificial intelligence (AI)-powered safety management agent, known as AI Risk Assessment (AIR), to small and medium-sized enterprises (SMEs) without charge. This strategic move is intended to democratize access to advanced safety technology, allowing smaller businesses, which often lag in AI adoption, to benefit from GS Group's expertise in AI transformation and industrial safety. The primary goal is to elevate safety standards across various industries by enabling SMEs to implement AI-driven safety protocols.
AIR was developed by GS Power, a subsidiary of GS Group, based on practical, on-site needs identified by field staff. The system integrates AI into job risk assessments, significantly enhancing both efficiency and quality in the process. This innovation has been recognized as a model case in industrial safety and earned a ministerial award from the Ministry of Employment and Labor for promoting a safety culture. The system functions by analyzing risks associated with industrial tasks using generative AI; users input a task name and description, and AIR automatically generates work processes, identifies potential hazards, assesses risk levels, and suggests preventive safety measures. This process, which previously involved manual reviews of health and safety regulations, can now be completed in approximately three minutes. Furthermore, AIR standardizes the quality of evaluations, which previously varied with individual expertise, and drastically reduces repetitive paperwork. The system's development emerged from a group-wide hackathon in 2024, where five GS Power employees specializing in safety and mechanical engineering built the system using GS Group's AX platform, MISO, without requiring code. GS Power has been using AIR internally since August, and this new initiative extends its benefits to a wider array of businesses.
Summarized using AI, subject to mistakes
Loading...