Scale and Organization of the Deepfake Crisis in South Korea
Investigations into Telegram networks in South Korea have uncovered a highly organized and systematic production of sexual deepfakes targeting students, teachers, and soldiers. Journalists and activists identified hundreds of chat groups dedicated to generating non-consensual deepfake imagery, often referred to as humiliation rooms or friend-of-friend rooms. These groups operate with strict entry requirements, demanding users submit multiple photos of individuals along with personal details such as names, ages, and locations. In some cases, dedicated rooms were created for specific students if sufficient content was generated, including rooms for underage pupils with over 2,000 members in at least one instance.
The crisis extends beyond universities, with middle and high schools also targeted. By the end of the investigation period, over 500 schools and universities had been identified as affected, though the exact number of victims remains unclear. A significant portion of the victims are believed to be under 16, South Korea’s age of consent. The perpetrators, many of whom are teenagers themselves, exploit AI-powered nudification bots to rapidly convert submitted photos into sexually explicit deepfakes within seconds.
The systematic nature of these operations was highlighted by journalist Ko Narin, who discovered groups where users actively solicited photos of classmates, colleagues, and acquaintances.
Impact on Victims and Societal Response
The psychological and emotional toll on victims has been severe. Heejin, a university student who received deepfakes of herself, described feeling petrified and isolated, a sentiment shared by many victims across the country. The revelation of the crisis’s scale has exacerbated anxiety, with victims questioning whether their online behavior contributed to the abuse. Some, like Ah-eun, have deleted their social media accounts entirely to avoid further exploitation, while others blame themselves for posting photos publicly.
In some cases, victims were told their cases were difficult to prosecute due to the fake nature of the images, despite the clear harm caused. The Advocacy Centre for Online Sexual Abuse Victims (ACOSAV) reported a dramatic increase in cases, counseling 86 teenage victims in 2023 and 238 in the first eight months of 2024 alone. In the week following the scandal’s exposure, an additional 64 teen victims came forward, overwhelming support services.
The crisis has also triggered a broader societal reckoning. Women’s rights activists argue that the deepfake epidemic is part of a long-standing pattern of online misogyny in South Korea, including spy cam abuse and verbal harassment. Structural sexism, they contend, enables such crimes to persist. Activists have criticized President Yoon Suk Yeol’s administration for downplaying gender inequality, cutting funding for victim support, and dismantling the government’s gender equality ministry. They demand systemic change, including stricter regulations on platforms like Telegram and mandatory education for young men on digital consent.
Legal and Platform Accountability
South Korea’s response to the crisis has been swift but contentious. Within a month of the scandal breaking, the country criminalized the possession and viewing of sexual deepfakes, enacting what is now considered the world’s toughest deepfake law. Police opened hundreds of investigations, and Telegram faced pressure to cooperate after threats of sanctions. The platform’s founder, Pavel Durov, was charged in France for enabling child pornography distribution on Telegram, adding international scrutiny to the issue.
Telegram has defended its moderation efforts, stating it uses AI tools and user reports to remove millions of violating posts daily. However, critics argue the platform’s light-touch moderation has allowed criminal activity to flourish. In 2019, a similar sex ring using Telegram ignored seven police requests for assistance before the ringleader was sentenced to over 40 years in prison. No action was taken against the platform at the time, raising concerns about repeated inaction.
Counselors and activists stress the need for comprehensive education on digital ethics, particularly among teenagers, where deepfake creation is sometimes trivialized as a game or prank.
Root Causes and Long-Term Challenges
Underlying the deepfake crisis are deeper cultural and technological issues. Counselors like Lee Myung-hwa note that deepfake abuse is normalized among teenagers, often treated as harmless entertainment.
The crisis also reflects broader societal attitudes toward gender and technology. Women’s rights groups argue that misogyny is the root cause, enabled by a lack of accountability for platforms and perpetrators. Without addressing structural inequalities, activists warn that similar crises will persist, exacerbated by advancing AI capabilities.
Possible Root Cause
The root cause of this crisis is multifaceted, combining technological, cultural, and systemic failures. At its core, the proliferation of sexual deepfakes on Telegram stems from the platform’s lax moderation policies, which have historically allowed criminal networks to operate with impunity. The ease of accessing AI nudification tools—often shared openly in these groups—has democratized the creation of non-consensual content, turning it into a scalable, industrialized form of abuse.
Culturally, South Korea’s deep-rooted gender disparities and the normalization of online harassment have created an environment where victims are often blamed for their exploitation. The trivialization of deepfake creation among teenagers, framed as a joke or prank, further desensitizes perpetrators to the severe harm caused. Additionally, the country’s previous failure to hold platforms like Telegram accountable—despite prior incidents such as the 2019 sex ring—has emboldened similar activities to continue unchecked.
While South Korea’s swift criminalization of deepfake possession is notable, enforcement remains inconsistent, and victims often face dismissive responses from law enforcement. Without addressing these underlying issues, the crisis is likely to recur, evolving alongside advancements in AI technology.