[Yoo Choon-sik] AI is no cure-all for Korea's economy

It has been more than three years since the use of generative artificial intelligence tools such as ChatGPT became widespread in everyday life, and they have already become indispensable for many workers. Using artificial intelligence to draft reports or analyze complex data is no longer uncommon.

According to a Bank of Korea report published earlier this month, 51.8 percent of workers in South Korea were using generative AI for their jobs as of 2025. While the pace of adoption is eight times as fast as the spread of the internet in the past, official data shows little meaningful improvement in the nationwide productivity over the past three years.

The disconnect is surprising because AI has enabled workers to complete the same amount of work more quickly, while often producing output of higher quality. The Bank of Korea report, therefore, examined whether AI has actually reduced working hours, whether saved time has translated into productivity gains and what conditions are needed to realize AI’s full benefits.

The analysis found that workers using generative AI spent an average of 3.8 percent less time completing the same tasks. Based on a standard 40-hour workweek, that translates into roughly 1.5 hours saved each week, and if all of that time were redirected toward production, overall productivity could rise by about 1 percentage point.

In theory, workers should have used those 1.5 saved hours to accomplish more work. However, the survey found a correlation coefficient of zero between reductions in working time and increases in output, in a phenomenon the authors described as an “AI productivity disconnect.”

The authors identified four main reasons for this outcome. First, AI remains concentrated on some very specific tasks rather than entire workflows, and only 4.4 percent of tasks experienced time savings of more than 20 percent.

The second factor is that introducing AI alone is unlikely to deliver meaningful performance gains without broader adjustments in corporate culture, employee behavior and work processes. This helps explain why productivity gains have been more visible among self-employed workers and professionals with greater autonomy than at conglomerates with rigid structures.

The third factor is that even if much of a workflow becomes more efficient, bottlenecks at specific stages can still delay the overall process. The fourth is that when rewards for additional performance are weak, workers may have little incentive to reinvest their newly available time into productive activities.

The report’s implications for the South Korean economy are fairly clear. Standardized tasks such as report summarization and data organization should be redesigned around AI-centered workflows so that saved time can be redirected toward higher-value activities.

By contrast, in open-ended work such as new business development and research and development, where human judgment and creativity remain central, AI should serve as a supporting tool rather than a replacement. Policymakers should focus on strengthening human capabilities while integrating AI into these areas.

The authors also stressed the need to redesign learning pathways so that junior employees do not lose critical opportunities to build skills by delegating basic tasks to AI. Still, these recommendations appear easier said than done when viewed against South Korea’s poor record of prompt and rational policymaking.

Even recent developments suggest that the report is presenting a formidable challenge to policymakers, politicians, economists and business leaders alike. While its conclusions may seem straightforward on the surface, implementing them within South Korea’s political culture and policymaking traditions is likely to be far more difficult.

International organizations and global investors have long argued that South Korea’s corporate culture and labor market practices suffer from deep structural weaknesses. Nevertheless, successive governments and major political parties have repeatedly promised sweeping reforms, only to fall short after taking power.

Although the current administration has been in office for just over a year, its policymaking record has generated as much concern as optimism. In particular, both the process and substance of revisions to the Commercial Act and labor-related laws and practices have revealed elements that deserve praise as well as scrutiny.

The Commercial Act revisions have been welcomed domestically as measures to address the "Korea discount" and strengthen minority shareholder protections. However, many international investors and foreign companies have warned that certain provisions could reduce predictability in corporate management, while the legislative process itself appeared vulnerable to criticism that outcomes had been predetermined.

The Lee Jae Myung administration’s labor policies are widely viewed as moving closer to a European-style model that strengthens labor rights. As a result, they have been welcomed by labor groups, progressive constituencies and environmental, social and governance-focused investors, while raising concerns among businesses, foreign investors and export-oriented manufacturers about competitiveness and labor market flexibility.

Supporters of the government’s rapid legislative approach argue that it is better to amend laws quickly and revise them again later than to prolong debate indefinitely. However, while perfect institutional reform is impossible in a single attempt, political considerations should not dominate efforts to address long-standing structural issues such as corporate governance and labor relations.

Both the Yoon Suk Yeol and Lee Jae Myung administrations have unveiled ambitious plans to embrace AI to restore momentum to South Korea’s otherwise stagnant economy. While such initiatives are timely and welcome, their impact will remain limited unless the country also tackles deeply rooted structural weaknesses in its corporate and social systems.

Recent media reports said that the Lee administration may appoint former presidential AI secretary Ha Jung-woo as the full-time vice chair of the Presidential Council on National Artificial Intelligence Strategy.

His experience at Naver Cloud and the presidential office makes the choice appear reasonable, but questions about his earlier resignation from the presidential office and unsuccessful parliamentary bid could cast doubt on the government’s commitment to breaking with past political practices if the appointment is confirmed.

Yoo Choon-sik

Yoo Choon-sik worked for nearly 30 years at Reuters, including as chief Korea economics correspondent, and briefly as a business strategy consultant. The views expressed here are the writer’s own. — Ed.

khnews@heraldcorp.com