Staff Reporter | Song Jianan
Recently, quite a few web novel authors have reported that Qidian Chinese Net has centrally dealt with or pulled down a batch of web novels with overly obvious AI traces, and a large number of hit works with over 10,000 monthly subscriptions have not escaped the “surgery” either. This move has sparked a lot of discussion in the web novel industry.
A post on social platforms shows that an author with 3,700 monthly subscriptions had their work penalized after the platform determined the AI-generated content ratio was too high. The author described the feeling as “being freed, like a fugitive finally getting caught.” Relying on AI to achieve decent subscription results while constantly worrying about platform audits and accountability—this is a common dilemma many web novel authors face.
On August 19, we reached out to Yuewen Group, the parent company of Qidian Chinese Net, for confirmation. An internal insider said that in order to maintain Qidian’s original content ecosystem, the platform’s governance over AI-written novels is an ongoing effort. They also clarified that the platform hasn’t actually taken down these works, but rather suggested that authors adjust their writing content.
We also learned that for the works being penalized, the books themselves can still be updated normally, and existing readers can continue reading and paying for subscriptions. However, they get directly removed from the two core leaderboards—the Monthly Ticket Chart and the Bestseller Chart—while also losing official key recommendation traffic and no longer receiving distribution opportunities to new platform users.
In Qidian’s product mechanics, the Monthly Ticket Chart and the Bestseller Chart are the most important gateways for new books to gain exposure. Rankings directly determine whether a work can land homepage recommendations and featured spots—they’re the core channel for new books to break through existing reader circles and attract a surge of new subscriptions.
Once kicked off the charts, a work loses its path to acquiring new readers and can only rely on existing loyal fans to sustain revenue. Over time, this inevitably leads to a downward spiral in earnings.
As for what percentage of AI writing triggers penalties, the platform hasn’t given a clear standard yet. Some web novel authors say that using AI for light polishing or error correction is fine, but the platform won’t cut special deals based on an author’s fame or past track record. For instance, high-profile works like “Mountain and River Chronicle,” “Striving to Be a Demon Lord Today,” and “Transcending in the Martial World” haven’t been spared either.
We checked Qidian Chinese Net and found that “Mountain and River Chronicle” boasts high subscriptions of 87,000 and monthly subscriptions of 32,000—definitely a top-tier hit—but it’s nowhere to be found on the platform’s various leaderboards. “Striving to Be a Demon Lord Today” and “Transcending in the Martial World” are also absent from the charts.
As early as April this year, Qidian Chinese Net issued a notice to authors titled “Announcement on Further Strengthening the Management of Non-Human Automated Creation.” The notice stated that to maintain a healthy original content ecosystem for online literature, the platform opposes using AI or other automated means to replace real humans in core content production or disrupt normal creative order, and will further strengthen standardized management of non-human automated text generation.
The notice emphasized that for verified non-human automated batch text generation, the platform will take measures including but not limited to pausing recommendations, removing from leaderboards, blocking works, and taking down works, depending on the severity of the violation. It will also publicly announce the handling results of typical violations on a monthly basis. At the same time, the platform will continue upgrading its AI detection and recognition capabilities, taking concrete actions to provide authors with a fair creative environment.
“We believe that good stories are precious crystallizations of human emotion and wisdom. Technological progress should become wings for authors, not a replacement for creation,” Qidian Chinese Net stated.
Qidian’s move isn’t an isolated case. Several mainstream web novel platforms in China are also rolling out AI content governance rules. Just today, Fanqie Novel under ByteDance released its July governance results, publicly sharing data and cases from the past month on cracking down on AI-generated shoddy content, padded writing, plagiarism, and other violations.
The announcement shows that in July, Fanqie Novel targeted prominent issues like AI batch generation, mechanical writing, meaningless padding, and chaotic formatting, strictly controlling the influx of low-quality content at the source of creation. The platform rigorously intercepted at the signing stage, refusing to sign a total of 160,000 AI-generated or hollow, padded low-quality books. It also cleaned up existing violative content in bulk, handling nearly 54,000 low-quality AI or padded books, and penalized 5,621 accounts involved in mass-producing low-quality content—all part of the fight against the content chaos caused by AI slop.
According to the platform’s governance details, the problematic works removed this time commonly suffer from mechanical and stiff writing, templated language, and chaotic formatting—all of which severely drag down the reading experience. Many works containing AI-generated low-quality content have been fully taken down. Meanwhile, the platform’s investigations found that some creators relied on AI to batch-produce manuscripts, resulting in disconnected storylines, logically confused plots, and collapsed narrative structures. These works completely failed to meet web novel creation standards and were all forcibly removed.
Back in April, 16 major online literature platforms including Yuewen Group, Fanqie Novel, Jinjiang Literature City, iReader Technology, and Qimao Novel jointly released the “Network Literature Industry Anti-Article-Laundering Self-Discipline Convention.” The convention explicitly opposes any form of article laundering, requires platforms to establish usage standards for AI-assisted creation, and makes clear that technological applications must be premised on respecting originality.
Facing the industry-wide crackdown, most original authors have voiced support, believing that AI-generated padded content is squeezing the living space of human authors and that cleaning up opportunistic works can improve the competitive environment. However, some mid-tier and newer authors are worried about vague audit standards, fearing wrongful penalties with no avenue for appeal. Meanwhile, some studios that rely on AI to cut costs and speed up production are facing the real pressure of having to restructure their business models.
Technical tools are neither right nor wrong in themselves, but the web novel paid-read model is built on the underlying logic of monetizing human creativity—good content is the foundation. How to balance the intensity of governance with avoiding collateral damage to legitimate creation is a question the entire industry will need to keep answering.
