On May 29, the YouTube channel “전과자” — known for bringing K-pop idols like EXO’s Kai and Chanyeol onto college campuses for content — deleted a video and issued a public apology after web novel readers and writers called out its use of AI-generated illustrations and an apparently AI-altered book cover to introduce a hit web novel. It was a narrow, specific apology. But it landed squarely inside a much bigger argument that’s been building across Korea’s web novel platforms all year: what happens once readers can tell, or just suspect, that AI touched the story they’re paying to read (Ilgan Sports, reporter Park Se-yeon, May 29, 2026).
Why One Apology Touched a Nerve
The video in question, titled “If an idol wrote fan fiction at a web novel department,” followed Kai and Chanyeol visiting Chungkang College of Cultural Industries’ web novel writing program. In the segment introducing novelist Baek Deok-su’s hit serial “I Have to Go to Work Even After Falling Into a Horror Story” (괴담에 떨어져도 출근을 해야 하는구나), the channel used AI-generated illustrations as visuals — and, critics noticed, what looked like an AI-altered version of the cover art from Baek’s other novel, “I’ll Die If I Don’t Debut” (데뷔 못 하면 죽는 병 걸림), as a thumbnail (Ilgan Sports, May 29, 2026).
The backlash was immediate enough that the channel pulled the video the next day. In its apology, “전과자” said it had been “lacking in respect for the web novel ecosystem and its creators, and in awareness of copyright,” and promised more care going forward (Ilgan Sports, May 29, 2026). For a single deleted video, that’s a fairly complete admission of guilt — and a sign of how little tolerance the web novel community currently has for undisclosed AI use anywhere near its work.
The Industry-Wide Dilemma Behind It
That low tolerance has a specific cause. Per Money Today, Korea’s web novel market — built around Naver-affiliated Munpia and Naver Series and Kakao-affiliated KakaoPage — runs on a structure where writers have to post anywhere from a few thousand to roughly 10,000 characters a day just to hold their ranking. That production pace makes AI assistance, for research, proofreading, or expanding scenes, “a temptation that’s hard to refuse” (Money Today, reporter Kim Pyeong-hwa, March 11, 2026).
The catch is what happens when it’s discovered. Money Today reports that when a writer’s AI use leaks through — an AI-typical phrasing, or a stray prompt response left in the text — readers respond with “rating-bombing” (별점 테러), and some serials have been suspended entirely after the backlash. Industry sources told the outlet that webtoon art makes AI use relatively easy to spot, but web novels are pure text, so there’s almost no way to detect it; platform review processes are currently built around violence and explicit content, not AI screening. Munpia is the one platform that has drawn a clear line, banning generative AI from its writing contests — Naver Series and KakaoPage have not adopted formal AI-use rules of their own (Money Today, March 11, 2026).
None of that resolves the underlying tension. AI assistance can genuinely help writers keep pace with the platforms’ relentless output demands — but if every writer leans on it at once, the production advantage cancels out, supply increases, and per-episode earnings come under more pressure, not less, according to Money Today’s industry sourcing. The “전과자” apology was a small, specific incident. But it’s a preview of how this argument keeps surfacing in public: not as a policy debate, but as a sudden, reputation-costly callout that platforms and creators alike haven’t figured out how to get ahead of.
Sources: Ilgan Sports (일간스포츠), Park Se-yeon, “‘전과자’, AI로 웹소설과 소개했다 논란…영상 삭제 공식사과 [왓IS],” May 29, 2026; Money Today (머니투데이), Kim Pyeong-hwa, “‘뭐야 AI로 썼어?’ 들킨 순간 별점 테러…웹소설 작가들 딜레마,” March 11, 2026.
