Corporate AI adoption rises 6.5-fold in 7 years, but gap between large and small firms widens fourfold

The share of South Korean companies using artificial intelligence rose 6.5-fold over seven years, climbing from 1.4 percent in 2017 to 9.2 percent in 2024 — but the gap between large and small firms widened sharply over the same period. The AI adoption rate difference between companies with 300 or more employees and those with fewer than 100 expanded from 2.8 percentage points in 2017 to 11.8 percentage points in 2024, a more than fourfold increase.

The Korea Research Institute for Vocational Education and Training published a report Thursday examining these trends, titled "How Far Are Companies Using AI? — The Spread of AI Adoption and the Gap Between Firms as Seen Through the Business Activity Survey." Researchers analyzed firm-level panel data from the Ministry of Statistics' Business Activity Survey covering 2017 through 2024.

The proportion of AI-using companies rose gradually in the early years — 1.4 percent in 2017, 2.7 percent in 2018, 3.1 percent in 2019, 3.4 percent in 2020, 3.8 percent in 2021 and 4.3 percent in 2022 — before accelerating sharply to 6.1 percent in 2023 and 9.2 percent in 2024.

The researchers said the acceleration "coincides precisely with the period when generative AI went mainstream in late 2022, triggering an explosive surge in public interest and accessibility." They added that changes in the external technology environment are being reflected in corporate adoption decisions relatively quickly.

The gap between companies of different sizes, however, is stark.

In 2024, the AI adoption rate among large companies with 300 or more full-time employees stood at 18.0 percent. Firms with 100 to 299 employees came in at 8.8 percent, while those with fewer than 100 workers were limited to 6.2 percent. Back in 2017, the differences were modest — 3.5 percent for large firms, 1.2 percent for mid-sized firms and 0.7 percent for smaller ones. As large companies pulled ahead, the gap between firms with 300 or more employees and those with fewer than 100 widened from 2.8 percentage points to 11.8 percentage points, a more than fourfold expansion.

"Large companies with abundant resources and personnel can more easily absorb the upfront costs and risks of AI adoption, secure specialized talent and leverage accumulated data," the researchers said. "Small and medium-sized enterprises, by contrast, face a triple constraint: adoption costs, lack of information and a shortage of skilled workers."

By industry, the information and communications sector posted the highest AI adoption rate at 34.5 percent — 3.7 times the overall average of 9.2 percent. Education services came next at 26.5 percent, followed by finance and insurance at 21.3 percent, electricity and gas at 15.5 percent, professional, scientific and technical services at 11.6 percent, wholesale and retail trade at 7.1 percent, business facility and support services at 6.4 percent, and manufacturing at 5.7 percent.

On manufacturing's low adoption rate, the researchers said it raised "concerns that a core industry accounting for a large share of employment could fall behind in the digital transition," while also calling it "a signal of opportunity, as there is significant potential to layer AI on top of already-established digital foundations such as smart factories."

Regional disparities were also evident. In 2024, companies in the Greater Seoul metropolitan area posted an AI adoption rate of 11.8 percent, 2.6 times that of firms outside the capital region at 4.6 percent. Among individual cities and provinces, Seoul led at 15.1 percent, well above the national average. Daejeon followed at 9.0 percent and Gyeonggi Province at 8.5 percent, with all other regions coming in below 7.0 percent.

Differences were also apparent in firm size indicators. In 2024, the median revenue of AI-adopting companies was 58.24 billion won ($41.3 million), 1.4 times that of non-adopting firms at 40.2 billion won. The median number of employees at AI-using companies was 178, 1.6 times the 111 recorded at firms not using AI.

The researchers called for support measures that "lower the entry threshold itself — such as AI vouchers, specialized consulting and shared industry-level infrastructure." For manufacturing, which has lagged in adoption, they said linking existing smart factory projects with AI deployment "would be an effective approach, using the digital foundations already in place as a lever."