Bank of Korea Analyzes AI Adoption and Productivity Effects
AI Use at Work Rises From Last Year... Paid Subscriptions Become More Widespread
Work Hours Saved Also Increase... From 1.5 to 3.7 Hours per Week
"Productivity Gains Remain Limited... Quantitative Expansion Must Lead to Qualitative Transformation"

More than six in 10 workers in Korea use generative artificial intelligence (AI), such as ChatGPT, at work, a survey found. The usage rate has risen from just over half last year, and personal paid subscriptions are also becoming more widespread. The time saved at work has increased from 1.5 hours to 3.7 hours per week. However, the effect of those time savings on actual work output or quality remains limited. This has led to calls for companies to rethink their work practices so that AI use translates into actual business results.

"Six in 10 Workers Use AI at Work... Saving 3.7 Hours per Week" View original image

The Bank of Korea (BOK) reported these findings in a BOK Issue Note released on the 17th, titled "AI Adoption and Productivity Effects: More Widespread, More Efficient, but Still Shallow" (Seo Dongwon, Oh Samil and Yoon Jongwon). Following last year’s survey, the BOK surveyed 16,342 employed people about their use of AI in August and September this year. This year, it expanded the survey and added questions. It also surveyed 2,024 respondents from last year again to build a dataset that could track changes over time.


AI Use at Work Rises from 51.8% to 66.5%; Weekly Usage Time Also Increases to 8 Hours

The analysis found that 66.5% of respondents used generative AI at work. This was up from 51.8% last year and the highest rate among major economies, including the United States and European countries. The average time workers used AI at work also increased, from 5.9 hours per week last year to 8.0 hours this year. In other words, they used AI for 22.4% of their actual weekly working hours.


"Six in 10 Workers Use AI at Work... Saving 3.7 Hours per Week" View original image

AI usage was higher among people in their 20s (77.1%), 30s (78.4%) and 40s (69.4%) than among those aged 50 or older (53.7%). By education level, usage was higher among university graduates (74.5%) and graduate degree holders (86%) than among people with less than a university degree (51.2%). By occupation, usage was high among professionals, managers and office workers.


"Generative AI has become more widespread, and the proportion of workers who use it regularly has also increased significantly," said Oh Samil, head of the Employment Research Team at the BOK's Research Department. "However, its use for core tasks remains relatively low, indicating that it has not yet been deeply integrated into the essential tasks of workers' jobs."


Personal paid subscriptions are also becoming more widespread, beyond subscriptions provided by employers. The survey found that 28.4% of workers who used AI at work paid for their own subscriptions, spending an average of 49,000 won per month. Based on this figure, annual subscription fees for all workers are estimated at approximately 2.8 trillion won.


Economic Effects: 3.7 Hours Saved per Week, but Productivity Gains Remain Limited

"Six in 10 Workers Use AI at Work... Saving 3.7 Hours per Week" View original image

Workers saved more time at work than they did last year. The average rate of time saved rose 2.4-fold, from 3.8% last year to 9.2% this year. Based on a 40-hour workweek, this represents an increase from 1.5 hours to 3.7 hours, saving nearly half a day. The time-saving effect was relatively greater among highly educated workers (4.9%) and paid subscribers (11.4%). If the time saved were reinvested in production, the potential productivity gain was estimated at up to 2.6%.


However, the report noted that the effect of time saved at work on productivity, such as increased work output or improved quality, remains limited. A closer look at the relationship between time saved and increased work output found that improvements in individual efficiency did not directly lead to higher output. The correlation between the two variables rose slightly, from 0 last year to 0.08 this year, but remained low.


The correlation between time saved and improved work quality was also low, at 0.09, indicating that workers who completed tasks more quickly did not necessarily see greater improvements in quality. Respondents said they used the time saved through AI to handle core tasks (33.8%) or take on additional work (29.3%).



"Although many respondents said they put the time saved back into their work, the weak relationship between the rate of time saved and the rate of increase in work output suggests that it is also important to focus on how those gains are converted into results after time is reallocated," Oh said. "This means that changes to work operations, including improvements to workflows, organizational restructuring and workforce redeployment, are needed to turn individual efficiency gains into increased organizational output." 


This content was produced with the assistance of AI translation services.

© The Asia Business Daily. All rights reserved. Unauthorized AI training and use prohibited.

Today’s Briefing