Futures
Access hundreds of perpetual contracts
TradFi
Gold
One platform for global traditional assets
Options
Hot
Trade European-style vanilla options
Unified Account
Maximize your capital efficiency
Demo Trading
Introduction to Futures Trading
Learn the basics of futures trading
Futures Events
Join events to earn rewards
Demo Trading
Use virtual funds to practice risk-free trading
Launch
CandyDrop
Collect candies to earn airdrops
Launchpool
Quick staking, earn potential new tokens
HODLer Airdrop
Hold GT and get massive airdrops for free
Launchpad
Be early to the next big token project
Alpha Points
Trade on-chain assets and earn airdrops
Futures Points
Earn futures points and claim airdrop rewards
AI plagiarism risks entering educational materials: Exclusive interview with Liu Liangcheng, winner of the Mao Dun Literature Prize: The style we develop over a lifetime should not be so disregarded and harmed by technology!
The Paper reporter | Ding Zhouyang intern Song Changzixing The Paper editor | Wei Guanhong
Recently, the winner of the Mao Dun Literature Prize and chairperson of the Xinjiang Writers Association, Liu Liangcheng, received an authorization request via the Chinese Text Copyright Association. A publishing house plans to include one of his articles in extracurricular reading materials for middle school students and hopes to obtain authorization.
“Once I read the title, I felt it was unfamiliar. I asked them to send me the original text, and when I saw the first sentence, I knew it wasn’t my article—it was an AI-generated imitation. ” Liu Liangcheng recalled to a reporter from The Daily Economic News, “The purpose of generating imitation with AI might be for traffic. But I think the root cause is that AI companies use writers’ content extensively to train large models. Using content that is in the public domain is of course fine. But for writers’ works that are still within the copyright protection period, AI uses them extensively without authorization, which is a huge harm to writers. A writer takes almost a lifetime to form a distinctive writing style and an imaginative world. When AI takes that for free as ‘training material,’ it will generate an overflow of imitation pieces. In the end, the content generated by AI imitation may even bury the author’s own content.”
“Writers’ rights have been harmed by piracy since the 1990s. Pirated books used to exist in bookstores and on street stalls, and even now some major internet platforms still sell them in large quantities. I didn’t expect that the harm from pirated books wouldn’t end, and then came AI stealing writers’ styles as well.” Liu Liangcheng said.
Regarding the “AI imitation incident,” on April 3, Liu Liangcheng accepted an interview with reporters from The Daily Economic News (hereinafter referred to as “The Paper’s reporters”).
Liu Liangcheng Picture source: provided by the interviewee
It’s a situation that’s both ridiculous and heartwarming: the winner of the Mao Dun Literature Prize has no choice but to personally “debunk” the AI-imitated text on his own social media account.
Once Liu Liangcheng’s post went up, within just one day it sparked a huge discussion on Xiaohongshu. “Oh my God, have you written an essay about medicinal herbs? I just did a Chinese language assignment based on it last week!” A netizen linked it to their own experience, “I think the article has been processed by AI. When doing the questions, I felt something was off—the questions and the article are completely disconnected.”
In response to this high-traffic “debunking post,” Liu Liangcheng told The Paper’s reporter: “It shows that copyright infringement cases involving AI-imitated texts are becoming more and more common. I often see people online promoting my books, and when I click in, I find it isn’t my article. Or it might be a caption on a video, a ‘golden quote’ with my name attached—similar to me in content and style, but also not my wording. And it’s impossible to trace the source.”
Liu Liangcheng then promptly showed The Paper’s reporter several online links related to imitations of his style. This means that in an information flow dominated by algorithms, the creators themselves may also become passive recipients of their own “digital counterfeit,” seeing with their own eyes how their style is dismantled and mimicked by technology, and then reappears in front of them under the name of an “original work.”
He also sent The Paper’s reporter a book review article about his work written by a “book blogger.” “It’s clearly written by AI, because the detail in my article that this book blogger praises—something they analyze—was never written by me at all.”
“In the last grouped discussion at the full committee meeting of the China Writers Association, I already put forward suggestions on this issue.” Liu Liangcheng believes that in the current AI era, it is urgently necessary to use legislative measures to build legal barriers for literary works. The development of artificial intelligence is, in essence, about absorbing the literary essence accumulated by humankind over thousands of years as its core training data. “For classic works that have entered the public domain, if they are used as humanity’s shared cultural heritage, of course we welcome it. But for the hard-won efforts of contemporary writers whose works are still under copyright protection, we must clearly and firmly advocate for legal protection. If AI wants to use them, it needs to obtain the writers’ authorization and approval and also pay fees.”
“A writer almost uses a lifetime to form their own unique writing style and imagined world. Yet AI, without any authorization from the writer, easily converts this wisdom crystallized through individual life experience into free training data, and mass-produces an enormous number of homogeneous imitations. This is a disregard for and harm to writers’ labor achievements. Even more worrying is that this harm will further expand. Because AI-imitated texts are in the process of becoming widespread and there is a lack of effective regulation, over time, the online space may be taken over by AI-generated ‘look-alikes,’ while the real content authored by writers themselves will be drowned out by the imitated content.”
As one netizen commented, “It feels like AI is like a fake person in a horror movie… slowly eating away and trying to disguise itself as the real person.”
Liu Liangcheng’s life trajectory and literary style are inseparably connected to rural Xinjiang. He once said that he has read countless books, “but the one thing that most influences my writing might be Xinjiang.”
Precisely because of such unique life experience and insight, in 1998 Liu Liangcheng published the prose collection One Person’s Village, shaking up the literary scene; in August 2023, after returning to rural life for ten years, Liu Liangcheng’s novel Benba won the 11th Mao Dun Literature Prize.
In the minds of many young readers, he is still the “mentor” of the Xinjiang writer Li Juan. When Li Juan submitted her work, someone suspected that Li Juan had plagiarized. Liu Liangcheng responded decisively: “Who is she going to plagiarize? There’s no such template in Chinese literature for her to plagiarize.” “I believe that land will produce its own speakers, just like it grows wheat and corn.”
This is also why Liu Liangcheng believes literary writing must be a craft. “Craft means using your own life to generate your own words.” Liu Liangcheng told The Paper’s reporter, “You must be a real experiencer of life in a place—having truly touched life, having gone through it. In your heart you have your own real pains, your own real sorrows and joys, and your own ups and downs. AI can’t and shouldn’t replace you to feel the real world.”
In this era where lectures and forums inevitably talk about AI, Liu Liangcheng also shared his views on AI at a readers’ conference not long ago. “I think the invention of this tool is timely.” Liu Liangcheng said to The Paper’s reporter, “All of the tools humanity ever had, from the Stone Age to the Bronze Age and the Iron Age, from horse-drawn carriages to internal combustion engines and to digital technology—every tool was based on extracting from the outside, taking from nature. Only AI tools are based on extracting from the knowledge that humanity has accumulated over millions of years. Each of us can’t possibly exhaust even the tip of the iceberg of human wisdom in a lifetime. At this time, we need a tool that digs inward and learns quickly. Nothing is more useful than AI.”
“I’ve also used tools from different stages of human development. I used pliers and even drove tractors. When AI tools came, I readily accepted it too. This tool is not the problem. The problem is how to use it.” In Liu Liangcheng’s view, “It’s impossible that once there are airplanes, we still only sit in horse-drawn carriages. But airplanes also have flight routes, and they can’t fly all over the sky. So human beings have always needed to regulate the tools they invent.”
In the field of literature, what counts as “fair use” of AI?
In Liu Liangcheng’s view, first, the use of writers’ works to train large models mentioned above needs to receive authorization from the writer; second, authors themselves should not abuse AI either.
“Now the editor of a literature publication is most worried about publishing an original article submitted by an author, only for it to be confirmed by the outside world as AI-generated. For editors, it is essentially a disaster.” He told The Paper’s reporter, “In fact, unless it’s the author themselves, it’s very difficult to identify AI-generated literary content.”
For example, the article that Liu Liangcheng intercepted—one that nearly made it into teaching materials—was edited by a professional, too. But because they wanted to find a simple and easy “essay reference” from a well-known writer, they failed to recognize it as a “fake text.” “The more famous the writer, the more serious the AI imitation.” Liu Liangcheng said.
On the other hand, this kind of thing—using AI for literary creation—is also common. “Especially in poetry, AI’s writing ability in poetry is quite strong. If it’s classical-style poetry, modern people may even find it hard to match, because we are far less trained in this area, while the ‘training material’ AI got from ancient poetry is most fully supplied.”
“I’ve also used AI when writing event planning proposals. When I encounter something I don’t understand, I will also use AI to learn quickly. But I suggest that when creating literature—especially for young authors who aspire to become writers—during the writing process, it’s best not to use AI, including using AI to make outlines, come up with titles, or draft text. Don’t use it for any of those. It will make you lazy, and it will also hinder your thinking.”
In Liu Liangcheng’s view, just like some things can’t be replaced by machines, literary writing should preserve its “handmade” and “craftsmanship” qualities.
Millions of news, precise interpretation—available on the Sina Finance APP
Editor: Shi Xiuzhen SF183