The adoption rate of Generative AI among small and midsize enterprises lags far behind large corporations, but an analysis found that when conditions such as company support are the same as in large corporations, the usage gap is not large. It argued that conditions need to be established quickly to close the AI usage gap between large corporations and small and midsize enterprises.
On the 10th, the Korea Chamber of Commerce and Industry Economic Research Institute said in its report, "Large-small enterprise gaps in Generative AI use: the role of capability and organizational environment," that the simple usage-rate gap for Generative AI between large corporations and small and midsize enterprises was 13.8 percentage points. It said 66.5% of large corporations and 52.7% of small and midsize enterprises use Generative AI. The findings are based on a survey of about 3,000 wage workers nationwide ages 20 and older.
However, when other factors such as company support systems and workers' prompt engineering capabilities were included in the analysis, the pure usage-rate gap attributable to company size itself fell to around 4 percentage points. The report said, "This suggests that even small and midsize enterprises can use AI as well as large corporations if an organizational environment for use is created."
In particular, when a company actively encourages AI use in-house, workers' AI use was 15.5 percentage points higher than in corporations that did not. Even when the company provided subsidies such as subscription fees, the usage rate rose by 8.1 percentage points. Workers' individual prompt engineering capability (23.5 percentage points) and acceptance attitude (21.4–40.0 percentage points) were also key factors boosting utilization. Prompt engineering is designing the questions and instructions given to AI.
These differences in support environments are clearly reflected in the reality of small and midsize enterprises' weak AI support infrastructure. A survey of Generative AI policies and work environments in corporations found that 70.4% of small and midsize enterprises did not have a Generative AI adoption roadmap. That was 16 percentage points higher than the 54.4% for large corporations, showing a significant gap in systematic AI strategies at small and midsize enterprises.
In most categories of company support, small and midsize enterprises fell far short of large corporations, including education and training (34.7% for large corporations, 24.9% for small and midsize enterprises), provision of internal guidelines and manuals (33.8% for large corporations, 24.3% for small and midsize enterprises), and provision of in-house developed or customized AI tools (11.4% for large corporations, 5.7% for small and midsize enterprises).
How time saved by Generative AI is used also differed between large corporations and small and midsize enterprises. Workers at both large corporations and small and midsize enterprises picked "investing in improving the quality of existing work" as their top choice for the time saved by AI. But for the second choice, large-corporation workers picked "undertaking new projects and tasks (22.6%)," while small and midsize enterprise workers chose "rest and personal time outside of work (27.3%)."
Research fellow Kim Yong-mi said, "A difference is observed between large and small enterprises in reinvesting time saved by Generative AI into creating new added value," and added, "Further research is needed, but this suggests that a short-term AI usage-rate gap could accumulate into a productivity gap in the mid to long term."
Polarization by industry and region was also clear. The usage-rate gap between large and small enterprises in services was 9.2 percentage points, while the gap in manufacturing reached 24.2 percentage points, or 2.6 times higher. Breaking down small and midsize enterprise usage rates by region, the Seoul metropolitan area (57.3%) far outpaced non-capital regions (47.8%).
The Korea Chamber of Commerce and Industry Economic Research Institute stressed that a comprehensive response by corporations and the government is needed to bridge the Generative AI usage gap between large and small enterprises. It suggested expanding AI-specialized courses within employment insurance vocational training to strengthen worker capabilities, and promoting tailored programs for blind spots such as non-capital regions and manufacturing. It said diagnosis and consulting as well as a standard roadmap should be disseminated so small and midsize enterprises can establish systematic adoption strategies, and that requirements for supporting AI subscription fees and tool adoption expense should be simplified to improve accessibility.