K-pop fans are well-known as a force to be reckoned with on the internet, with some fans holding fierce allegiances and passions when it comes to specific artists.. On social media, some obsessive fans are using generative AI to create what is essentially self-insert fan-fiction\u2014creepy videos and images of themselves kissing, cuddling and getting various kinds of lovey-dovey with their favorite idols. The result is a contemporary, tech-powered mutation of idol culture that takes fan art to new extremes, and many people in K-pop fan communities are understandably pretty upset about it.Users on Reddit and various K-pop forums have expressed outrage at these deepfake fantasies, calling on other fans to report those who use AI tools to generate nonconsensual videos of their K-Pop heartthrobs. In one thread on the K-pop Uncensored subreddit, a user posted redacted screenshots from an AI-generated video by another fan showing themselves being hugged and kissed on the cheek by Keonho from the boy group CORTIS, who is a minor. \u201cSome fans may think it’s harmless but if you tolerate this, you’re opening doors to much worse things in the future,\u201d Reddit user wbu_lizzie wrote in the thread. \u201cThe idols didn’t consent to this, they are real people living their lives, not a bunch of wattpad \/ ao3 characters.\u201d\u201cPeople have been cutting and pasting these kind[s] of things forever but the more realistic the images get the more dicey it gets psychologically,\u201d wrote user Goldie_Prawn, in another thread on K-pop Uncensored calling out AI usage. \u201cI think even with traditional photoshop the work and skill required gives some level of mental buffer to the creator that isn’t there with AI.\u201d<\/p>\n
Dr. Sarah Keith, a professor at MacQuarie University in Australia who has done extensive research on K-pop fandom, says that the AI deepfake videos come from a long legacy of fan-created content that goes far beyond the world of K-pop. Fan fiction and art have always been an essential component of fan culture, and while these creations often skirt the boundaries of copyright, they are generally not viewed as creepy or exploitative. But with the advent of generative AI tools that can quickly create realistic images and videos of real people, Silicon Valley has built a technical workaround for the law and the social norms of consent.\u201cK-pop fans (like all pop culture fans) have long been active producers of fan materials about their preferred idols. However, these to date have not been ‘convincing’ as artefacts that might have involved the idol themselves,\u201d Keith told 404 Media. \u201cIt’s a parallel issue involving both fans and AI tools. Because of these two factors, there is now an increased facility for fans to create large volumes of content, which do not respect the idols’ personal integrity, and which can be distributed online in ways that undermine the idols’ reputation (e.g. sexualized deepfakes).\u201dWhile many fans spoke out against these creepy behaviors, others pointed the blame at the K-pop industry\u2019s business model that\u00a0encourages parasocial fandom\u00a0 while exploiting many artists and fans alike.\u201cThis is a sad side effect of an industry that for years [has] pushed parasocial relationships on their fans,\u201d wrote Serious-Wish4868, in a thread on r\/kpopthoughts about the AI-powered sexualization of idols. \u201cWhen idols and companies are constantly pushing you to live out your romantic and sexual fantasies with idols in everything they do from social media posts to live interactions, what do [you] expect to happen?\u201dBut some of the entertainment agencies that manage K-pop artists, called companies\u2014and the artists themselves\u2014have expressed they\u2019re not happy about the situation either. Members of hugely popular K-pop groups like BTS and Tomorrow X Together have spoken out against the use of generative AI to create digital simulacra using artists\u2019 likenesses. In March, OA Entertainment, a label founded by and representing Blackpink member Jennie, warned in a press statement that it would begin taking \u201cstrong legal action, including civil and criminal proceedings, against any acts that infringe upon the artist\u2019s reputation, rights, portrait rights, and other intellectual property rights.\u201d The statement came in response to harassment and stalking incidents involving the artists, and heavily implied the issue of AI-generated content without naming it explicitly.At the same time, many of these idol talent companies are embracing generative AI tools as a vehicle for cutting costs throughout the industry. SM Entertainment, one of the biggest labels in Korea which represents major acts like Girls Generation and Super Junior, has described generative AI integration as a core part of its business strategy, and last year it released a fully AI-generated music video for the girl group Aespa. And last December, the K-pop world saw the launch of a fully synthetic idol group called GLXE (pronounced \u201cgalaxy\u201d), which combines AI-generated songs with uncanny AI avatars.\u00a0\u201cThey’d automate the whole industry if they could make money,\u201d Lindsey Knuth, a K-pop fan and associate editor at Tangle News, told 404 Media. \u201cPornographic deepfakes of minor idols is an issue, as is the industry’s (lack of) response, to me.\u201dThe cultural tolerance for fan-generated AI slop varies depending on region and context. South Korea currently ranks number one in consumption of AI slop content, as the country\u2019s tech industry pushes adoption and cultural acceptance of the technology. In the US, however, polling consistently shows that the majority of young people have extremely negative opinions of AI tools and AI-generated content\u2014and are becoming more and more anti-AI as jobs, schools, and institutions pressure or force them to use it.Keith notes that while labels and talent companies have always been very active in protecting artists and their reputations, they only do so from the perspective of protecting an investment. Fans, on the other hand, generally tend to be more emotionally invested and likely to defend idols as human beings, both from creepy fans and from the companies themselves.And like with AI and tech more generally, the lack of technical or legal safeguards against the technology\u2019s many well-documented harms has made social shaming one of the only effective deterrents against abusive behavior in the K-pop world.\u201cIt’s reassuring that many fans are openly calling out non-consensual uses of AI and deepfakes, because in the absence of technical guardrails on this kind of content, social acceptability is really what’s important,\u201d said Keith. \u201cSo yes, it is reassuring that many active fans online are building a culture of respect where idols’ likenesses are not used in non-consensual ways.\u201d<\/p>\n","protected":false},"excerpt":{"rendered":"