On June 26, 2026, Researcher Kyoungmi (Kimmy) Oh appeared on SBS News Hunters' debate segment "Issue Sniper." There, she delivered opposing views on the mandatory AI filtering of illegal filming footage and the law regarding manipulated false information. This post compiles the full questions and answers that were shortened or left out during the live broadcast due to limited time.
Q1. Starting in July, large websites and online communities will be mandated to run AI filtering systems to block pornography and illegal filming footage (under the Telecommunications Business Act). What is your overall assessment ahead of the implementation of this system?
A. There is absolutely no disagreement regarding the urgency of eradicating illegal filming footage and protecting victims. I firmly believe such crimes must be rooted out. At the same time, however, it is true that I have concerns about whether the means to achieve this goal are justifiable and whether they cross constitutional boundaries.
Q2. What are your thoughts on the concern that this system constitutes "prior censorship"?
A. There are certainly people who might feel uncomfortable or resistant to the civil society's argument that this is prior censorship. However, I believe it does amount to prior censorship. The objective of the system—to block or eradicate illegal acts—is undoubtedly well-intentioned. Even so, a structure that mandates a filtering system to monitor private telecommunications in real-time, driven by an administrative agency rather than the judiciary, inherently aligns with the definition and essence of prior censorship, which is prohibited by our Constitution.
To use an analogy, imagine if the government introduced a system where an administrative agency inspects the bags of every single citizen daily just to catch thieves. Even if the police or administrative agency claims they are only checking whether a bag contains an item identical to a stolen one, wouldn't we still consider that an infringement on citizens' rights?
Q2-1. There are also concerns that once this measure takes root, regulations could expand into other areas. What is your take on this?
A. This is precisely the point that civil society organizations are worried about. They are concerned that the administrative agency's judgments could lean in a pro-government direction, leading to the expansion of regulations into entirely different domains.
In fact, most of these measures originate from the goal of creating a "safe and healthy society." During the pandemic, our government tracked the movements of confirmed patients for the sake of a "safe society." Because these movements were disclosed in extreme detail early on, it led to severe side effects, including privacy infringements and the stigmatization of minority groups.
Even when starting with good intentions, these systems can bring about unexpected consequences. This is why any system that restricts fundamental rights must be designed and introduced with utmost caution.
Q3. Some might wonder, "If an AI does it, isn't it exempt from being called censorship?" What difference do you see there?
A. Just because artificial intelligence delegates the screening process does not change the fact that the government is conducting the screening. Automated immigration control at airports is done by AI, but no one thinks it isn't a state-run inspection.
Q4. Above all, can we trust the criteria of an AI that judges right from wrong? How can we verify and accept it? Although a registered database (DB) exists, isn't it true that even platform operators don't know exactly what is inside that DB?
A. It is true that there have been many errors. Since the scope of application is set to expand to still images, even more errors may occur. The problem is that the authorities seem to be forcing the use of AI to reduce these errors. On the internet, countless people use platforms where others build communities to mediate information. If the number of users grows, platforms are forced to equip expensive AI systems. This inevitably oppresses freedom of expression as a whole.
Furthermore, there are valid doubts regarding the filtering technology itself. If a system is to be implemented, there should at least be discussions on who manages the database, how it is managed, and how often it is updated. However, there is hardly any questioning raised about this, and the information is not being disclosed to the public.
Q5. This law was amended in June 2020 following the "Nth Room" incident. Yet, critics point out that while illegal filming footage is systematically distributed on sites with foreign servers, regulations are heavily concentrated on mainstream domestic platforms like Naver, Kakao, and DC Inside, as well as domestic users. How effective is this system?
A. Unfortunately, I don't think it will be highly effective. One of the reasons people upload illegal filming footage is to profit by selling them on the dark web or similar spaces. Therefore, it is difficult to see real effectiveness by regulating mainstream platforms.
Rather than imposing mandatory regulations, we should utilize an intermediary liability exemption system so that information intermediaries can voluntarily delete and block such content freely. Europe's DSA (Digital Services Act) and the US's DMCA (Digital Millennium Copyright Act) utilize intermediary liability exemptions, achieving significant results where hundreds of millions of pieces of illegal filming footage are voluntarily blocked by intermediaries each year.
Q6. Some argue that the fundamental solution is to drastically increase the severity of punishment for digital sex offenders so that they fear distribution itself. Despite this, is there still a reason why materials must be blocked in advance?
A. Considering the perspective of those who demand prior blocking, I believe the severity of punishment for digital sex crimes has indeed become higher than before. This increase was clearly an achievement brought about by the women's rights movement. However, when comparing it with other laws, it would realistically be difficult to raise the level of punishment much further.
Moreover, due to the nature of digital content, once illegal filming footage is exposed to the internet, it can spread uncontrollably, and completely deleting it is close to impossible. I believe this is why the necessity for prior blocking is being so strongly pushed forward.
Q7. The "Amended Information and Communications Network Act," which imposes an obligation on large platforms to block false or manipulated information, is also taking effect. Outfits supported by government-affiliated organizations will judge what qualifies as such. How do you evaluate this structure?
A. A system where an administrative agency judges whether an online post is illegal or not is a highly outdated practice. People's opinions all start from false assumptions. Even news articles from media outlets do not convey facts exclusively; they convey opinions as well. I believe the very ambition to clearly discern falsehood from truth is an unachievable desire.
Q8. If a report regarding "fake news or false information" is filed on a large platform, the platform must delete the post or suspend the account. Failing to do so can lead to "criminal punishment." Do you see this as a foolproof mechanism to stop distribution, or a measure that chokes the platforms?
A. European countries do not have administrative agencies judge illegality. Instead, they have strengthened the responsibility of each platform to activate reporting and blocking systems, while also reinforcing mechanisms through which the affected party whose post was blocked can file an objection. As a result, the number of posts blocked due to reports increased massively, reportedly reaching hundreds of millions. Yet, cases requesting restoration were said to be very few.
Not only that, but they have also built and strengthened fact-checking systems led by civil society organizations and researchers. The structure is set up so that society itself—rather than an administrative agency—judges the illegality and harmfulness of information.
Korea, on the other hand, has chosen the method of imposing liability on platforms through rigid regulations. This administrative-convenience approach can actually reduce the incentives for platforms to voluntarily strengthen their accountability.
Civil society organizations oppose state censorship; they do not oppose platforms blocking or deleting illegal filming footage according to their own terms of service. Perhaps the DB that platforms could build on their own might be far more extensive than the DB provided by the Korea Communications Commission for cross-checking.
Providing policies that encourage platforms to devise their own methods (such as technological or policy development) to build independent databases and filter out illegal filming footage more thoroughly might be a better way to create a healthier internet ecosystem. If the state forces a specific method like this, platforms might become more passive in technology or policy development, thinking, "I did everything the government told me to do, so I won't put in any more effort."
Technology and policies require continuous updates. Shouldn't we also consider cases where the technology or policies distributed by the state fail to keep up with reality?
Q9. The most hotly debated part is the "5x punitive damages system," and opinions for and against it are sharply divided. What is your view on this?
A. Our country does not have a general punitive damages system. A person who commits a malicious act faces no risk of paying punitive damages, yet a person reporting on that malicious act faces the risk of paying five times the damages if they happen to report inaccurate facts. This carries a high risk of shrinking not only the media's role of criticism and surveillance but also civic consciousness as a whole.
Q10. The internet space we desire is a safe place where one does not become a victim of digital crimes, while at the same time, a place where one's privacy and freedom are not diminished. Can we catch these two rabbits with a realistic alternative? Please give us your closing remarks.
A. Currently, there is no alternative technology or suitable workaround that can perfectly block illegal filming footage. Therefore, completely eliminating digital crime from this world is impossible.
Nevertheless, the logic that we should use means that deface constitutional values just because there is no alternative is dangerous. Rather, precisely because there is no perfect alternative, I believe our society needs to engage in more fierce dialogue to find ways to minimize the infringement of fundamental rights while lowering crime rates and strengthening ex-post (after-the-fact) measures.
Investigating and punishing crimes after they occur, deleting illegal filming footage, and robustly establishing support systems to help victims recover are all vital. A comprehensive approach is required. It seems necessary to reflect on how much effort our society has actually put into building such a comprehensive system.
[Related Posts]
[Open Talks] Open Net Appears on SBS News Hunters to Oppose Mandatory AI Filtering Systems (2026.06.29.)
0 Comments