16 prevention projects target 92,835 cases this year against 393,789 flagged sites
Park Jeong calls for stronger staffing and closer alignment with AI predictions
An AI-powered prediction model operated by the Korea Occupational Safety and Health Agency has flagged more than 390,000 workplaces nationwide as high-risk, yet the agency's annual targets for on-site accident prevention guidance and inspections this year cover only about 92,000 cases — prompting calls to expand the staffing and organizational capacity needed to translate risk identification into actual field support.
According to data submitted to National Assembly lawmaker Park Jeong, a member of the Climate, Energy, Environment and Labor Committee, the agency's high-risk workplace prediction system classified 393,789 sites — or 12.78 percent of the roughly 3.08 million workplaces nationwide — as high-risk.
By contrast, the combined annual target for 16 accident-prevention guidance and inspection projects the agency is pursuing this year — including risk-assessment consulting, technical support for workplaces where serious injuries have occurred, and patrol inspections — stands at 92,835 cases. That amounts to just 23.6 percent of the number of high-risk workplaces identified.
However, the figures are not directly comparable, as target units vary across projects — some count visits or support sessions while others count individual workplaces. A closer look is needed to determine whether the same workplaces are receiving support multiple times and how many of the flagged high-risk sites are actually being reached.
Annual targets by major project are: risk-assessment consulting and certification, 20,000 cases; patrol inspections, 20,000 cases; region-specific programs, 12,500 cases; chemical exposure information outreach, 8,500 cases; and rapid technical support for workplaces where serious injuries have occurred, 7,200 sites.
A gap also emerged between the prediction results and the actual targets of prevention projects. According to Park's office, the agency's model shows that the proportion of high-risk workplaces rises with company size, yet projects targeting small and medium-sized businesses account for 51,200 cases — about 55 percent of this year's total prevention targets.
Park called on the agency to more actively incorporate risk prediction results into the selection of prevention project targets, arguing that having identified high-risk workplaces, the agency should clearly define which sites receive support and expand the scale of field prevention activities.
"Now that the system is in place, the agency must expand its prevention programs and sharpen its targeting so that the work translates into actual accident prevention," Park said. "To that end, it needs to build up the necessary organization and personnel, and actively reflect the prediction model's results in selecting targets for prevention projects."
fact0514@heraldcorp.com
