The growing use of generative AI to create content for commercial purposes is leading to more cases of consumer misguidance and is now raising concerns about a regulatory gap in Korea’s AI law.
Under the Framework Act on the Development of Artificial Intelligence and the Creation of a Foundation for Trust, which took effect in January, businesses that use generative AI to produce and sell content are not clearly distinguished from ordinary AI users.
Accordingly, they are not explicitly subject to transparency requirements such as disclosure obligations.
This regulatory blind spot came to attention after consumers complained about the mobile app Book Thief, which offers Korean translations of classic literature whose copyrights have expired.
Related Article
On July 25, users flooded X with complaints after the app’s operator disclosed — more than a year after launching the service — that its translations had been generated using AI.
The app had advertised that users could permanently own 1,000 translated works of public-domain classics for 19,800 won ($14).
“We created this app so people with limited budgets could enjoy reading at a time when books have become expensive,” a representative of Book Thief said in a post on X on July 25, following the criticism.
It added that although the translations were generated by AI, they were reviewed by humans, so it could guarantee their quality.
Many users, however, responded that they would not have purchased the service had they known the translations were AI-generated.
Book Thief shut down its service six days after posting the statement and is processing refund requests.
The use of AI-generated images in product marketing is even more widespread.
Some sellers use AI-generated models in product pages to showcase cosmetics or clothing, while others use AI-generated figures dressed as doctors, pharmacists or other professionals to recommend products in advertisements.
As similar cases have increased, the Fair Trade Commission revised its guidelines for the review of endorsement and testimonial advertising in June, which is currently in effect.
The revised guidelines additionally stated that advertisements featuring endorsements by virtual characters may constitute unfair advertising if the endorsements do not reflect actual experience.
Businesses are increasingly turning to generative AI because it reduces marketing and promotional costs.
AI-generated models also eliminate the reputational risks associated with human celebrities, making them attractive to advertisers. Courses teaching businesses how to use generative AI for marketing are also being sold.
Despite the rapid growth of such practices, the AI framework act does not clearly define businesses that commercially use AI-generated content as AI service providers, leaving the boundary between AI service providers and commercial AI users unclear.
“Businesses that use AI could fall within the scope of the law, but it remains unclear whether a company that merely uses commercially available AI as a tool qualifies as an AI service provider under the act,” Lee Yong-hae, an attorney who serves as an adviser to the International Association for Artificial Intelligence and Ethics, said.
“As AI-based businesses continue to grow, the law should be revised to distinguish between ordinary AI users and businesses that use generative AI for commercial purposes.”
Major e-commerce platforms, including Naver Smart Store and Coupang, have introduced their own guidelines to fill the regulatory gap.
Naver Smart Store announced in June its AI-generated content guidelines for sellers. The rules took effect on July 10.
The platform recommends that sellers disclose when images or videos used in thumbnails or product pages have been created or altered using AI.
It also prohibits the use of AI-generated content for certain product categories, including medical devices, infant food and fresh food. Sellers who violate the guidelines may face penalties, including having their products removed from Naver Shopping search results.
“As the use of AI becomes more diverse, we established guidelines on AI-generated content in advance to provide a safer shopping environment and better protect consumers,” a Naver Smart Store official said.
Coupang also issued guidelines for sellers last month requiring promotional content created using generative AI to avoid false or exaggerated advertising.
“Depending on the context, AI-generated content can seriously mislead consumers [when they make purchasing decisions],” Choi Kyoung-jin, a law professor at Gachon University, said.
“Just as advertising disclosure guidelines were introduced for internet search results about a decade ago, it is now time to discuss whether regulations should be introduced for businesses that sell AI-generated products or use AI-generated images in commercial activities.”
