A plan to redraw along four tracks the workload of employment centers, which 2 million people pass through each year, was unveiled on the 29th of last month at the Korea Press Center in Jung-gu, Seoul. Its backbone is that repetitive administrative work is handed to artificial intelligence (AI) while staff concentrate on counseling that connects job seekers with companies and on career planning support. That means moving processing that runs by fixed rules, such as paying unemployment benefits, over to the machine side. Separating administrative organizations from service organizations, setting up an AI-based digital employment center, and building staff expertise are the tasks bundled here.
At the forum, prepared jointly by the Ministry of Employment and Labor and the Korea Labor Institute, it was pointed out that a single counselor handles more than 70 unemployment recognition cases a day. Unemployment recognition refers to the procedure in which a person receiving unemployment benefits has their job search activity verified. It is a volume that leaves barely ten minutes for one case.
The room for counseling narrows accordingly. Since the 2005 modernization of employment services, the burden at the counter has tilted toward unemployment benefits.
Vice Minister of Employment and Labor Kwon Chang-jun said in his congratulatory address that the essence of the innovation is to leave repetitive administration to AI and fill the space it frees with services that connect people and jobs.
Gil Hyun-jong, senior research fellow at the Korea Labor Institute, presented the "Plan for Innovating National Employment Service Delivery" and proposed pulling apart the administrative work organization and the service work organization. He also put forward a plan to set up new regional centers and have frontline centers put their weight on face-to-face services. "My Own AI Employment Center," which analyzes a job seeker's career and capabilities and picks out services that suit them, was also brought to the discussion table that day. Beneath it lies the idea of moving away from a counter that hands out unemployment benefits toward a lifelong total care system running from graduation to retirement.
The discussion also pointed to gaps in the design. Yoon Dong-yeul, professor at Konkuk University, noted that the classification of tasks has to be sorted out first, beginning with where to draw the line between benefit administration and employment support functions. He also added a call for concrete measures on how to move administrative staff freed by automation into counseling and case management.
Case management refers to a method of following and looking after a single job seeker all the way to employment. Kim Gyun, senior research fellow at the Korea Employment Information Service, proposed running the two axes at the same time rather than taking the design of the reform plan and the overhaul of laws in sequence.
The timing of the plan is also worth watching. An analysis found that over the past year the number of employed people aged 15-29 in information and communications and in professional, scientific and technical services fell by more than 80,000, the largest drop since the relevant statistics began in 2017. Even though growth rates rebounded in the first and second quarters of this year, the number of young employed people fell by more than 200,000.
The Korea Development Institute (KDI) warned that within the next 10 years more than 250,000 jobs a year could disappear in mid- and high-skilled occupations such as professionals and office workers. President Lee Jae-myung, putting forward a startup-centered nation, stressed a shift in employment strategy suited to the AI era.
Chatbots and automated screening have been running at private-sector counters for quite a while. Public employment services stand on different conditions. Unemployment benefits are money that livelihoods hang on, and how to keep open a channel for a person to look again when the screening goes wrong is the core task in designing automation. This plan is a case in which the state has stated that it will itself draw the boundary between automated work and counseling work.
The task of dividing organizations touches laws and the organizational structure. That is also where the proposal for parallel implementation comes from. What decides success or failure is where the freed-up time flows, more than how quickly automation arrives at the counter.
The results of the overhaul will be confirmed by how much the time counselors spend on counseling job seekers instead of processing documents actually increases. Only when that time grows at counters that 2 million people pass through each year does career planning support become possible.
