4 in 10 Koreans use generative AI: survey

The ChatGPT app icon on a smartphone / Reuters-Yonhap

Nearly four in 10 Koreans have used generative artificial intelligence (AI) such as ChatGPT and Gemini, a survey showed Thursday, but a persistent digital divide sidelines the majority due to a lack of digital literacy.

A survey by the Korea Media and Communications Commission (KMCC) showed that 38.9 percent of respondents have used generative AI. The figure is up 14.9 percentage points from 24 percent in 2024, according to the commission's 2025 User Panel Survey in Intelligent Information Society. The survey polled 4,324 internet users nationwide aged 18 to 72 who use smartphones at least once a day. Users spent an average of 49.6 minutes per day on the platforms.

A chart generated by artificial intelligence (AI) shows the most common uses of generative AI among Koreans.

The most common use for generative AI is information search, with 86 percent citing its efficiency. Meanwhile, 72.6 percent of users said using the technology saves time, followed by educational purposes at 68.2 percent, and solving complex problems and supporting daily tasks, both at 64.8 percent.

However, concerns over the risks of generative AI intensified, rising 6 to 10 percentage points from a year earlier. Fraud and criminal exploitation topped the list of fears at 67.7 percent, closely followed by job displacement at 67.6 percent, decreased creativity at 66.8 percent and the spread of misinformation at 66.4 percent.

Among non-users, 63.5 percent cited a steep learning curve or a lack of know-how as the primary technical barrier. Subscription costs deterred 59.2 percent, while 59.1 percent reported difficulty choosing a service. Another 59 percent flagged both privacy concerns and ethical issues.

The KMCC said the findings demonstrate an ongoing need for user support policies designed to close the access gap and improve AI literacy.

Facing these stark divides and growing risks, respondents overwhelmingly agreed on the need for strict corporate accountability. The survey showed that 81.3 percent see the need for generative AI providers to establish internal audits and reporting procedures to prevent users from generating outcomes that infringe on personal rights, while 79.3 percent demanded that recommendation service providers disclose their algorithmic criteria to ensure transparency.

This article from the Hankook Ilbo, the sister publication of The Korea Times, is translated by a generative AI system and edited by The Korea Times.