The spread of low-quality AI videos continues unabated…The AI ​​Basic Law is also insufficient

It is difficult to regulate videos that have been re-edited by removing the watermark.
Display obligations under the AI ​​Basic Act do not apply to users.

Low-quality videos created using generative artificial intelligence (AI), known as “AI slop” (slop, meaning “waste”), are rapidly going viral on major platforms like TikTok and YouTube. However, there is still no clear solution to effectively block this phenomenon. The platform aims to prevent the proliferation of AI-generated materials by enforcing monetization standards by mandating separate labels to indicate AI-generated content. However, concerns remain that both users and authors can still remove or manipulate AI-generated labels, leaving the issue a regulatory blind spot.

A June 23 analysis by U.S.-based video editing platform Kapwing found that 59% of the 10,742 videos published to new accounts on TikTok were identified as AI slops, three times the rate of 21% on YouTube.

The main reason for the popularity of AI slop is the revenue structure of video platforms. The major platforms make money based on the number of views and watch time, not the quality of the video. Creating videos using generative AI reduces production costs, making it easier to make money by feeding low-quality videos in large quantities.

Platform companies aren’t sitting idly by as AI becomes more widespread. Starting in 2024, YouTube will require creators to disclose whether their content was generated or modified using AI, and in May began applying an automated system to detect and label the use of AI with its own AI detection technology. Starting from the 29th, Naver will strengthen the monetization standards for its short-form service Clip, aiming to increase the proportion of high-quality videos. Content will be screened for the use of AI and mass-produced materials, and advertising incentives will not be applied to low-quality content.

However, the issue of video creators changing AI-generated labels during the secondary editing process remains a loophole. As a result, AI-generated materials with labels removed or altered are redistributed on online platforms, making it difficult to verify their authenticity. This year, South Korea enacted an AI basic law that requires the labeling of AI-generated content, but the regulation is limited to companies, making it difficult to resolve the issue. It remains difficult to prevent distributors and users of AI-generated materials from reprocessing or damaging these labels.

Baek Jeong-min, a professor at the Department of Media Studies at Kyung Hee University, said, “Unless there are clear issues such as violation of portrait rights, explicitness, or violence, there are insufficient legal regulations governing individual video works.” He added, “These are side effects of the falling cost of AI-powered video production. Users need to develop media literacy skills to discern whether content is generated by AI or not.”

©Asian Economic Newspaper (www.asiae.co.kr) Unauthorized reproduction is prohibited.