Kim Min-ho, Chairperson of the Korea Internet Self-governance Organization (KISO) Special Committee for Review of false or manipulated information, holds a press briefing on the 2nd and presents the committee's activities and review status. /Courtesy of Shim Min-kwan

If each platform has different standards for determining false or manipulated information, it can burden users and cause social confusion. We judged that member companies of the Korea Internet Self-governance Organization (KISO), including Naver and Kakao, need to deliberate by applying the same standards.

Kim Min-ho, Chairperson of the KISO Special Committee on Deliberation of False or Manipulated Information, explained this background at a press briefing on the 2nd for how major domestic platforms, including Naver and Kakao, came to apply joint standards to false or manipulated information.

In the first deliberation applying the joint standards, not a single post was recognized as actual false or manipulated information. The committee finally deliberated 23 cases, including exaggerated information and user reviews related to products and services and posts suspected of being generated by artificial intelligence (AI), but all were decided as "not applicable."

The amended Act on Promotion of Information and Communications Network Utilization and Information Protection, which took effect on July 7, prohibits the distribution of false or manipulated information and allows large platforms that receive reports to take measures such as deletion, access blocking, and exposure restrictions. Unlike illegal information that the government directly deliberates, this is a self-regulatory area where platforms make the determination. To reduce differences in judgments by platform, KISO formed a special committee of experts and prepared common guidelines in June. Kim, the Chairperson, said, "Just because it is false information or manipulated information does not mean it is immediately subject to a distribution ban," adding, "There must be recognition of an intent to cause harm or a purpose to obtain unjust benefits, and an infringement of others' rights or the public interest."

◇ All 23 cases in the first deliberation "not applicable"… 70% of reports were reviews or promotional posts

It was expected that the first recognized case of false or manipulated information would emerge after the amended law took effect, but no posts were decided for measures such as deletion or access blocking. Initially, KISO expected that politically sensitive posts would be mainly reported. In reality, product and service information, user reviews, and promotional posts accounted for 19 cases, or 70% of all reports. Cases included using photos different from reality when introducing lodging facilities or exaggerating the efficacy of specific substances without verification.

Kim, the Chairperson, said, "Member companies brought to the committee cases among reported matters that were ambiguous to judge," adding, "So far there have been no cases where domestic platforms have indiscriminately deleted or blocked posts to the extent of infringing users' freedom of expression."

The decisions of the special committee have binding force under an agreement between KISO and its member companies, not legal compulsion. Once the committee determines whether something constitutes false or manipulated information, each platform decides the specific level of measures, such as deletion, exposure restriction, or displaying a warning message.

◇ The hardest judgment is the poster's "intent and purpose"

According to the guidelines, to determine something as false or manipulated information, all three requirements must be met: ▲ false or manipulated information ▲ a purpose to cause harm or obtain unjust benefits while knowing it is false or manipulated ▲ infringement of others' personality rights or property rights, or the public interest.

Kim, the Chairperson, singled out the poster's intent and purpose as the most difficult requirement to judge. Kim said, "We can judge whether specific infringement has occurred or whether there is a possibility of infringement, but it is difficult to verify the poster's subjective intent in their mind," adding, "We infer the purpose through objective facts such as the posting motive and the relationship with the victim, repetitiveness, and the method of dissemination."

Standards include whether the poster repeatedly targeted a specific victim, whether they intentionally removed watermarks or other marks for source verification, and whether they attempted organized dissemination using bots or macros. Not only economic profit but also social or political benefits such as expanding influence or increasing followers can be included in "unjust benefits." Falsity is not recognized merely because part of the information differs from the facts. The assessment considers whether it materially affected ordinary users' recognition of facts, in light of the core message and the overall context. Subjective opinions or value judgments, predictions and estimates, and academic, scientific, or religious debates where the truth is not settled are, in principle, excluded from judgment.

◇ AI content judged in context… Global big tech also discussing participation

Content generated or edited by AI does not automatically become manipulated information. The key is whether it has a "deceptive appearance" that could lead ordinary users to mistake it for objective facts. Even if it is labeled as AI-generated, if it lacks creativity and deceives users to produce specific benefits or harm, the judgment is made by comprehensively assessing the overall context.

Kim, the Chairperson, said, "Before the law took effect, there were strong concerns that the rule banning the distribution of false or manipulated information could significantly shrink freedom of expression," but added, "After applying the law's requirements one by one to actual cases, we are cautiously judging that it may not seriously infringe freedom of expression as vaguely feared."

KISO plans to expand the joint self-regulatory framework to overseas platforms as well. Kim, the Chairperson, said, "KISO membership is open without distinction between domestic and foreign," adding, "We are currently discussing membership with at least two global big tech corporations, and we will disclose it once confirmed."

Currently, a total of 17 companies, including Naver, Kakao, Daum, and Nate Communications, are participating as KISO member companies.

※ This article has been translated by AI. Share your feedback here.