"Huk, My Face in Porn?" The Evolving Crime of Acquaintance Defamation [Han Seung-gon's Case Notebook]

Using AI Technology to Synthesize Faces Not Only in Images but Also in Videos
Creating So-Called 'Jiin Neungyok Videos' and Indiscriminately Spreading Them on SNS
Victims Suffer Horrible Pain... Petition to Cheong for Stronger Punishment

A website that created pornographic videos by synthesizing the faces of Korean female idols.

[Asia Economy Reporter Han Seung-gon] #Choi Mo (female, 26), a university student, recently experienced a shocking incident. Her friends kept sending her KakaoTalk messages saying, "Isn't this you?" Upon checking, she found that her face had been superimposed and distributed in pornographic material. Choi naturally denied that the photo was of her, but questions from acquaintances continued. She said, "No matter how much I say it's not me, there seem to be people who don't believe me," adding, "I have reported it to the police and am also taking steps to have it deleted." She then expressed her frustration, saying, "I feel like I'm going crazy."

Recently, a man in his 30s who superimposed the faces of others or acquaintances onto pornographic videos and distributed them was caught by the police. This man committed the crime using artificial intelligence (AI) technology. The crime involved AI technology called 'deep learning.'

This crime is already a global issue and is called a 'deepfake crime' because it involves attaching the word 'fake.' It is a type of digital sex crime where one's face can be superimposed and distributed in pornographic videos anytime and anywhere.

According to the Cyber Investigation Unit of the Jeonbuk Provincial Police Agency on the 1st, the man is suspected of opening a so-called 'acquaintance humiliation room' on Twitter and distributing composite photos, including nude pictures, to about 1,200 followers.

From August 2017 to October this year, he used AI technology to meticulously superimpose the faces of celebrities or acquaintances onto specific videos containing sexual content. It was also revealed that he produced these fake videos targeting children and adolescents among his Twitter followers.

Given this situation, citizens expressed anxiety about the 'deepfake' crimes that have penetrated daily life. Lee Mo, a university student in her 20s, said, "It's really terrifying to think that my face could be superimposed onto pornographic videos or other obscene materials," emphasizing, "I hope such crimes can be prevented technologically in advance."

Kim Mo, a company employee in his 30s, called for stronger punishment. Kim said, "We need to stop these crimes through severe punishment," adding, "Isn't it practically impossible to delete such content once it's distributed on Twitter or social media?" He expressed concern, saying, "In this regard, they should be severely punished. Victims must endure unimaginable pain."

The characteristic of the 'Jiin Neungyok' crime is that it is indiscriminately spread on social media platforms such as Twitter. From the victim's perspective, they cannot help but suffer extreme pain.
[Image source=Yonhap News]

As citizens' anxiety shows, this crime has already deeply penetrated our daily lives. According to the Broadcasting and Communications Review Committee's 'Focused Monitoring of Obscene and Prostitution Information' conducted in the second half of 2017, among the 494 cases detected and blocked, 291 cases involved acquaintance humiliation and composite images, the highest number.

'Acquaintance humiliation' refers to indiscriminately distributing online images where photos of people one knows, general public photos obtained from the internet, or celebrity photos are superimposed onto obscene materials, accompanied by personal information or defamatory sexual content. Applying AI technology allows such composites to be made in videos as well, causing significant social problems.

In fact, on a U.S. online community, illegal content was indiscriminately distributed, superimposing the faces of famous foreign actresses such as Scarlett Johansson and Emma Watson, as well as domestic celebrities like Seolhyun, onto pornographic videos and photos. Because these videos are meticulously edited using AI technology, it is difficult to distinguish their authenticity, causing victims to suffer terrible pain.

In this context, related crimes are on the rise. According to a report by the British BBC citing a deepfake research report from the cybersecurity company 'Deeptrace,' the number of deepfake videos, which was about 8,000 in December 2018, increased to 14,698 by December last year. This means the number nearly doubled in about a year. Among these videos, 96% are pornographic.

[Image source=Yonhap News]

Given this situation, in August 2018, the mother of a female student who was a victim of 'acquaintance humiliation' appealed on the Blue House's public petition board for a swift investigation and to share her daughter's suffering.

The petitioner, a mother raising a high school daughter, wrote, "I am writing this with a relieved heart after being shocked," and lamented, "Since the closure of Soranet, posts promoting obscene crimes have been actively circulating on Tumblr and foreign sites, and my child has become a victim."

She continued, "My daughter is an ordinary high school student, but her photo in a school uniform was cleverly edited and posted with posts promoting sexual assault and disgusting language. I am afraid for her safety on her way to school and don't know what to do," expressing her anxiety.

She urged, "Urgent deletion and strong punishment are urgent matters. The disclosure of my child's school name and grade causes her great emotional pain," calling for a prompt investigation.

There was also a petition asking to stop 'deepfake acquaintance humiliation' crimes. The petitioner explained, "Artificial intelligence uses deep learning and internet big data to superimpose a person's face onto other videos," lamenting, "Some Korean celebrities have already become victims, and word has spread across various domestic communities."

She added, "Access to Reddit and Tumblr, where all kinds of prostitution, nude composites, and deepfake materials circulate, should be banned domestically, and more police personnel and budgets should be allocated to investigate the distribution of deepfake and composite photos," emphasizing, "Deepfake is a clear sex crime and defamation that leaves shame on the face owner in the video."

Meanwhile, with the amendment of the 'Special Act on the Punishment of Sexual Crimes' in June this year, not only simple producers but also distributors of such digital sex crimes can be punished with imprisonment of up to five years or a fine of up to 50 million won. Those who commission production can also be punished as accomplices.