GuShiio.com Article source: Silicon-Based Research Laboratory
Image source: Generated by AI
After three years in the online literature circle, this was the first time that editor Huahua saw five or six AI submissions appearing in her mailbox in one day: “This year, AI manuscripts have been flooded.”
In the studio editorial group, the latest topic of discussion has also become “how to distinguish AI authors and AI manuscripts.” Some online editors collectively call these manuscripts “DS flavor too strong”-a subtle metaphor, a faint sense of human-machine sense There is also a sudden turning point. In addition to receiving manuscripts and signing contracts, online editors began to post “boycott AI manuscripts” on social media, calling for the return of pure land to the online literary circle.
Online editors dislike “AI manuscripts” Photo source: Xiaohongshu screenshot
At the same time, since the beginning of the New Year, online text platforms including Tomato, Jinjiang, LOFTER and other online text platforms have successively introduced judgments or penalties for “AI manuscripts.” While receiving the support of some online text authors, they have also caused many users to use AI-assisted writing. Newbie authors feel panic and anxiety. They do not know what the boundaries of using AI to write are and what the criteria for platform determination are.
“I just used AI to properly expand and polish the content, but traffic dropped sharply during the verification period.” Newbie author Qiqi told the “Silicon-Based Research Laboratory”. Qiqi said that many authors have also been “accidentally injured by platforms and soft-titled.” Some online writers have begun to worry that the “AI storm” will affect their own profits and decided to wait and see for a while and wait until this time passes before opening a new book.(Note: The “verification period” is the work recommendation mechanism of the Tomato Platform. When the author’s work is updated to 80,000 words after signing, it can enter the traffic pool to earn income. The platform provides a 7-day verification period, and after 7 days, the work will be updated to 100,000 words will enter the official recommendation period, which is the so-called debut show.)
This is not the first “struggle” between AI and the online cultural circle.
As early as July last year, the “Author Boycott Tomato Novel AI Agreement” became a hot topic on social media. Over the past year or so, as the forefront of AIGC’s implementation, the integration of online literature platforms and AI has continued to accelerate.
From the platform, author and editor, the confusion and confusion of the entire online literature circle can be regarded as a cruel and beautiful AI reality story-it seems to be an efficient productivity tool, but in the specific practice process, when a semi-familiar technology meets reality, it often faces a game of multiple parties.
Different from the previous round of online article writers boycotted AI, this round is the online article editors who lift the table.
Since December last year, Huahua has clearly discovered that there are more and more AI manuscripts in the mailbox: “Sometimes I can receive five or six articles a day, and there are even many second-source traffickers who submit them repeatedly.” There are generally three ways for online editors to identify whether it is an AI manuscript: The first is to use AI to identify AI,For example, using relevant AI Detection Tools such as Rosefinch, Doubao, Chinaz AI, and MitataAI,The second is manually assisted identification,Editors will identify the main content, style, personality and specific sentences based on their own experience;Third, the studio itself will also have corresponding measures,Types of AI drafts will be summarized, relevant AI authors will be blacklisted, etc.
But although there are many forms of determination, editors spend more effort communicating with the author. Generally speaking, for suspected AI manuscripts, the editor will communicate with the author first to avoid accidentally injuring the author. Online article authors can provide self-certification-such as codeword videos, document history records, etc.,”We usually don’t do one size fits all, and we communicate with the author first.” Huahua said.
But faced with the question “What exactly is an AI manuscript”, it is difficult for editors to give an accurate answer,”The software is biased, and I am also learning how to distinguish it.” Some online editors mentioned that after DeepSeek came out of the circle, it promoted AI’s writing style to become gorgeous and rich. AI became more and more aware of online text routines, which also made it difficult for editors to identify. In the process of communicating with authors, they even had to detail specific sentences and paragraphs: “For example, looking at the transition and the unnatural nature of people.” Huahua said.
Qi Mao editor Li Zhi gave more detailed criteria. AI was pieced together randomly. She found that a manuscript used the phrase “Hell is too cold, you come to die for me.” This sentence actually came from a Jinjiang classic from her early years. She repeatedly confirmed with the author whether it was written by herself, but the author said “yes.” On “self-certification”, Huahua mentioned that authors can provide chapter records, historical time, etc. of the code words and writing, and can even use previous drafts to compare the style of writing for editors.
However, the above-mentioned way of identifying AI is beyond the understanding of some new online writers. Qiqi mentioned that many online literature platforms have AI touch-up functions, and some authors also took their manuscripts to the testing platform and found that the testing AI generation probability is very high, which puts them in a situation of “difficulty in self-certification”.
Some new author also posted a complaint, and was commented by readers as “AI is too strong”, but she obviously did not use AI.”You can say that my writing is poor, but don’t say that my writing is like AI.”
As a result, the current online literature industry has fallen into an “AI cycle”–The platform encourages authors to use AI to assist writing. Authors boldly try to use AI, but have to use AI to test themselves to prove that they “are not AI.” Online editors add new workload due to AI writing.
It confuses editors and authors in the online literature circle and the platform’s attitude towards AI.
Judging from the content scale of some online text platforms alone, the magnitude of AI-generated content has increased exponentially.According to big online data, on March 1, 2024, the number of new books on Tomatoes debuted reached 5606, compared with 400 in the same period last year, a year-on-year surge of 1302%.In an interview with Red Star News, the team of Zhihu “Yan Yan Novels”, the head platform of short online articles, also said that the number of submissions of AI creation content has indeed increased significantly recently.
Online culture platforms are the first players to embrace the AI wave. As early as July 2023, Reading Group launched the industry’s first large model,”Reading Wonderful Pen” and writer assistant wonderful pen version. Zhihu cooperated with Wall Wall Intelligence to develop the “Zhihaitu AI” large model, targeting online literature supply and demand. On both sides, AI has been integrated into the entire process of production, distribution and experience.
After DeepSeek’s release, Reading also announced that its writer-assisted creative product, Writer Assistant, has integrated the independently deployed DeepSeek-R1. Reading said that it will greatly enhance the writer’s assistant’s question and answer reasoning ability and descriptive polishing effect. The creative model “Chinese Carefree” under Chinese Online also announced the deployment of DeepSeek. Take the Tomato Writer Assistant as an example. When you enter the book opening page and enter ideas, AI can provide creators with multiple inspiration references. During the specific writing process, the platform also provides an AI toolbox-including AI expansion, AI polishing, AI continuation, AI assistant, etc.
Tomato Writer Assistant AI Function Source: app screenshot
The positive attitude of online culture platforms towards AI is not difficult to understand–One is to reduce costs and increase efficiency, and the other is to optimize business models.
From the perspective of cost items, the cost expenditure of online text platforms can be divided into content costs, platform distribution costs, and film and television production costs (TV dramas, online dramas, movies, animations, etc.). Among them, the main content costs related to creators are content costs, involving writers ‘remuneration, copyright procurement, etc. Taking reading as an example, in the first half of 2024, content costs accounted for 20.4% of revenue, which means that reading must give nearly a quarter of the income to authors.
Source of reading cost item: financial report
However, AI can not only efficiently improve the efficiency of creators, but also expand production capacity. This is undoubtedly a multi-pronged achievement for readers.
In addition, AI can also explore more monetization methods and open up new space in member fees and IP copyright operations. For example, tomato novels can create a closed loop of “online text + short drama” business; the advantages of reading are in IP pricing and operation, you can use AI to feed back new tracks such as “IP+ games, IP+ going to sea”.
But as Huahua feels, the platform’s attitude towards AI seems to be positive, but when it comes to online creators, it is “watching while walking”:”The platform’s idea is to embrace and rectify.” Some online editors also mentioned that the creator’s attitude towards AI this round has actually been amplified: “What the platform bans are actually accounts that produce meaningless content in batches and malicious water. In fact, there have always been accounts, but now everyone is too sensitive to AI, so it may be amplified.” An online editor explained to us.
Jinjiang Literature City, which is regarded by the online literature circle as “the first shot against AI in online literature,” is not completely resisting AI. Bing Xin, director of Jinjiang Literature City, mentioned in an article on February 17 that AI writing assistance is still a new thing, academic discussions are insufficient, legal provisions are imperfect, and originality is difficult to define.”Whether AI is used to define” is far more difficult than originality.
Jinjiang Literature City released the “Trial Operation Announcement on the Use and Judgment of AI-Assisted Writing”
Source: Screenshot of Bishui Jiangting website
This is also the biggest dilemma for AI online literature at present:First, when AI intervenes in writing, how do we define what is the originality of a work? The second is how do platforms, authors and editors judge the “AI footprint” under user reports, that is, whether the author used AI?
At present, most of the judgment standards given by the platform focus on solving the latter problem. The platform will establish a “grading system” and establish red lines for AI writing assistance based on its own experience.
For example, Jinjiang is divided into textual assistance-type and creative assistance-type. Its Chinese typefaces are divided into proofreading level, description level, and narrative level according to different degrees. Creative type is divided into element level, rough outline level, and fine outline level. Jinjiang currently accepts three levels: textual assistance-proofreading level, creative assistance-element level, and creative assistance-rough outline level.Using AI to proofread text, assist outline generation elements, etc. are actually allowed by the platform.
“The wheel of history is coming faster than I imagined.” Bing Xin’s sigh has aroused the resonance of many people. For the online literature circle, AI is the general trend, and this shock will not be the last time. Many editors believe that the most important thing at this stage is that the platform launches targeted AI original definition standards, and the industry has jointly established a red line. At the same time, creators also need to adjust their mentality: “Currently, pure AI text is still easy to identify.” Huahua has repeatedly emphasized to us that some creators need not panic too much.
She feels that this round of AI discussions has put some creators and editors on opposite sides, which is what the editors don’t want to see the last thing:“The online literature circle has been difficult and exciting in recent years. We (editors) hope that the online literature ecosystem will develop in the positive direction more than anyone else.”
Technology develops first, but platforms and editors react later. This natural lag in technology has also supported a group of “shovels” behind them-both those who assist AI and those who are anti-AI are popular.
Waveform Intelligence, an AI startup that previously launched the AI-assisted creation tool “Frog Writing”, was acquired by OPPO last year. A company called GPTZero, established in January 2023, relies on “verifying AI content”. Revenue reaches millions of dollars in one and a half years.
We learned that in addition to the chatbot supported by large models, the common AI tools in the AI online literature circle also include some vertical applications, such as Xingyue writing, code mile writing, etc.
Source of related AI writing tools: App screenshots
The service provider told us that there are currently several “rules for making quick money” for AI online articles: “I suggest that you write a female video and earn about 300-800 yuan for earning full attendance. You can also sell abandoned manuscripts on different platforms.” When we asked “whether there is a risk of AI testing being blocked”, the service provider clearly said:“You’d better not sign a contract with an editor.” According to him, some platforms that receive short reports currently adopt automated review, and the probability of manual review is small, so it is easier to sign contracts and complete manuscripts.
According to data released by China Academy of Social Sciences, China’s online literature industry will usher in a market of 300 billion yuan in 2023. The number of online literature creators will be close to 25 million, and the number of online literature users will reach 537 million.On average, 4 out of 10 China people read online literature. This is of course a vast market for AI to gold.
In addition, the unique differentiated ecosystem of the online culture circle also determines the profitability of AI entrepreneurship.
First of all, online article creators are basically large and their needs are also differentiated, which means that AI online article creators can quickly iterate functions based on needs, find PMF, and quickly verify them.
Secondly, there is a strong paying propensity. For some novices and middle-waist creators, making money by submitting online articles is the king’s way. Therefore, entrepreneurs choose the right tool, coupled with the Internet’s “fission” gameplay, and they can quickly accumulate the next batch of loyal users in the private domain.
Wan Lei, head of Frog writing and co-founder of Waveform Intelligence, previously shared in an interview with “Silicon Star Man” that for novel authors, writing is a tool for making a living. They are unwilling to disclose their means of making money, and even hope that fewer people know better, so“Relying on private domain to polish products well” is also a good path.
Finally, there are the “people” among entrepreneurs. According to our understanding, in the AI online literature entrepreneurship, there are a group of “great gods” with rich experience in online literature writing and are very familiar with the workflow of online literature creation (such as popular online literature style, book opening, rough outline, detailed outline, expansion, etc.) To a certain extent, the online literature itself has more than 20 years of commercialization, and relevant skills, routines, and frameworks have been precipitated into “standards” and “templates”, which are easy to be AI.
As Huahua observed, in this round of AI turmoil, most of the people who use AI to write in batches are novices and AI writing studios. Their characteristics are that they leave after making a quick buck and don’t care about “raising pen names.” This is also the inevitable trend of the “Matthew effect” in the online literature circle-leading authors eat meat, and middle waist and bottom-level authors drink soup. As Bing Xin wrote in the forum, what the AI turmoil will not affect is leading creators: “The income of top entertainment authors has increased instead of falling because their works are” anti-AI replication-resistant.”
The fragmented ecology has also determined the current attitude of the online literature circle towards AI differentiation-a group of technological optimists have begun to try to write in AI. In order to deal with the so-called “AI flavor”, they have evolved “de-AI flavor prompt word reference”; there are also a group of conservatives who insist on “rubbing hands” and resisting AI manuscripts, which is an inevitable phenomenon during every technological iteration.
In an online writing group, the future of “AI writing” is still discussed late at night. Optimistic people say: “Counteracting tools are meaningless.” One creator issued a soul torture:“Will online literature be meaningless in the future?”
References:
1. Top of the list: If AI will definitely rewrite the online literature industry, please give me a way to write it
2. Red Star New Culture: Online editor, encirclement and suppression AI
3. Silicon Star Man: AI application entrepreneurs are desperately trying to grab traffic in the cracks of large factories
