Luo Chong is using a mainstream artificial intelligence large model software currently on the market to create, mainly using the two functions of Wenshengtu and Tusheng Video. First select a reference picture, and then write a prompt text to briefly describe it. Within two to three minutes, AI can quickly generate a 5 to 10 second video or animation.
His individual experience is echoing a larger need. All 90 places for courses related to AI comics and dramas in the "Skills Night School" in Zhenjiang, Jiangsu Province, were filled in one day, with students ranging from secondary school students to stay-at-home mothers to uncles in their sixties. Not long ago, at a youth employment fair in Chongqing, more than 200 emerging digital positions including AI short drama executive directors and AI editors were unveiled, attracting a large number of graduates to stop and inquire. Some companies said that there is a particular shortage of talents in this type of position, and the maximum monthly salary can reach 20,000 yuan.
The craze is spreading across the country. On September 17, at the press conference on the theme of "Getting Started with the 15th Five-Year Plan" held by the State Council Information Office, Wang Xiaoliang, Director of the Online Audiovisual Program Management Department of the State Administration of Radio and Television, introduced that in the first eight months of this year, 430,000 micro-short dramas were launched across the country, 13 times the number in the whole of last year, of which more than 90% were produced with AI assistance.
When AI significantly lowers the threshold for content production and professional skills, a new employment and entrepreneurial ecosystem around AI is accelerating to take shape. But under the tumult, there are both real money and silver gains, and there are also hidden traps.
Embrace the trend: one person is a crew
Luo Chong, 38 years old, studied drama, film and television directing in college. After working, he worked as a director at a TV station and as a publicity agent at a major Internet company. In the past two years, he has chosen a new career direction - AIGC (artificial intelligence generated content) short play director.
The first time he came into contact with AIGC technology was in 2023. He first looked for tutorials online, and then signed up for an offline training class. In 2024, he and several friends participated in the "Prompt the Future" Shanghai International AIGC Creation Competition, and won third place in the video group with the AI short film "Jiading Huilong". "Getting such results among more than 100 works gave everyone a great encouragement." Luo Chong said that at that moment, he realized that the future was coming and he should embrace the trend. Since then, this group of like-minded young people have worked side by side and made many different explorations.
With the help of AIGC technology, creators first generate pictures by inputting text descriptions of the pictures, and then generate videos from the pictures. This process sounds simple, but it is not easy in practice. Luo Chong compared it to "stir-frying", "if it's bland, add more salt, and if it's overcooked, learn to add less salt next time." Just as the seasoning needs to be just right, the image also needs to be beautifully balanced. In order to achieve the most satisfactory effect, creators need to constantly adjust scene, clothing, hair accessories, lighting and other picture elements to find the most harmonious "recipe" for the picture.
Therefore, behind every shot that the audience sees in front of the stage, there is 10 times the hard work, that is, there are 100 shots in front of the stage, and there are up to 1,000 shots in preparation and screening behind the scenes. To deal with such a huge workload, the "yield rate" of AIGC technical tools is particularly important, which directly determines production efficiency. The level of "yield rate" depends on the ability to accurately understand and reproduce people's movements, postures, expressions and text semantics in the real world. During creation, Luo Chong sometimes uses more than a dozen AIGC tools with different capabilities.
At present, the orders Luo Chong undertakes are mainly advertising and promotional videos. His team members are young creators from all over the world and from all walks of life, including documentary directors, programmers, designers, and mothers. They work together through remote online collaboration on a daily basis. "Many of them were not originally engaged in film and television-related careers, but everyone has a love for co-creation. Many members also have daily jobs. Everyone uses their spare time to discuss and improve to make one thing the best."
In the field of AI creation, being a director or being able to edit is no longer a hard threshold for entry into the industry. Learning ability and creative enthusiasm are more important. Chen Yongren (pseudonym) born in the 1990s is an example. He has sold more than 3,000 AI videos on the Xianyu platform. Before that, he worked in photography in Shanghai. In the first half of 2025, after paying attention to the market demand for AI videos, he quit his photography job and took orders for AI video production full-time. After receiving stable orders, Chen Yongren settled in Chengdu and established his own studio at the end of last year.
Luo Chong said frankly that in the past, when he was a director, he was responsible for arranging the work of crew members. Now he is giving instructions to the AI to "submit the Tao", and one person can become a crew. He took the AI short film "Porcelain Rhythm" as an example. "In the past, actual shooting required photography, lighting, stage management, director team, and scene notes. The total number of pre- and post-production staff would not be less than 30. Since it involves a large number of cross-regional transitions, even if it works day and night every day, it will take about a month to actually shoot." With AI, it can be completed in a week, and the production cost has been reduced to one-third.
The transformation is not only reflected in cost and efficiency, but also reshapes Luo Chong's way of working. He also built a studio at home. Compared with the days when he had to commute to and from work in a company and was cut off by meetings and processes, he now has more time for himself.
More and more entrepreneurs and creators are beginning to flock to this track. Qichacha data shows that as of October 9, there are 24,300 AI short drama-related companies in China. In terms of registration volume, the number of registrations in 2025 will reach 5,600, a year-on-year increase of 29.21%. So far, 4,426 related companies have been registered this year.
Some people earn more than 10,000 yuan a month, and some are considering turning their side jobs into regular jobs.
From photographer to AI video creator, from working alone to having his own team, Chen Yongren's transformation path is a microcosm of how more and more young people are realizing "light entrepreneurship" with the help of AI tools.
During the 2026 China International Fair for Trade in Services held in September, Xianyu released trend data related to "light entrepreneurship" in 2026: There are more than 1.5 million light entrepreneurs on the platform who have obtained stable income from continuing operations. In the past six months, AI transactions have been the fastest-growing segment of service transactions, with orders increasing by 157%. Among them, AI programming and AI comics-related services have the strongest demand, with the number of orders increasing 17 times and 14 times year-on-year respectively.
"More and more people are entering the industry, and the video quality required will become increasingly sophisticated." Chen Yongren said that among the opportunities in the market, there are also hidden challenges for every AI service seller. His order library includes both commercial video needs from large enterprises and small entrustments from a large number of individual buyers. For example, a boy once came to him, hoping to use AI to make a proposal video using photos of him and his girlfriend when they met and fell in love. There are also many middle-aged people who asked Chen Yongren to restore old photos of their deceased elders and make them into AI videos, so that the deceased elders could say a few more family wishes during their lifetime and keep them in their thoughts.
"The price depends entirely on the demand, and there is no unified pricing." Chen Yongren explained that individual buyers' simple video requirements are easy to make, require few modifications, and can even be completed in one hour, so the transaction can be completed for tens to hundreds of yuan. For commercial orders for large enterprises, the quotation will generally increase exponentially. For some commercial videos of large brands, due to their high production standards, the price per second of footage can be between 200 yuan and 300 yuan. Even so, compared with the traditional advertising industry, it is still cost-effective. Chen Yongren said that it is normal for traditional TVC advertisements to cost more than 100,000 yuan for a finished film, while AI production can compress the price of the same effect to 30% to 40% off the original price. This is one of the reasons why AI videos are favored by brands.
Binbin (pseudonym), who graduated from the China Academy of Art, has a main job in design and a side job in helping people train models on Xianyu. In 2022, when Binbin first heard about large model fine-tuning technology like LoRA, it was still a niche technology. In the past two years, large AI models have been soaring. Binbin couldn't sit still. He went to Station B to read tutorials and learn while taking orders. Firstly, through trading, he put pressure on himself to learn, and secondly, he could understand the latest directions and common technical difficulties through market demand.
Because of his art skills, Binbin will be more professional in designing-related AI model training. He said that converting professional terminology in the art field into language that AI can understand is also a science. From being unfamiliar to becoming proficient, Binbin now earns more than 1,000 yuan a day from his side job. He has even seriously considered turning model training into a formal entrepreneurial project.
However, embracing the trend requires more calmness. The HR of a cultural communication company told reporters that when faced with emerging positions related to AI, many people submit resumes based on the word "AI", but when asked if they can write storyboards and control the rhythm of the content, they cannot answer. She suggested that school students should make career plans in advance and use actual project experience to prove their ability to get started; students looking for jobs across majors should identify their own advantages and combine AI tools to form differentiated competitiveness. “There are many opportunities in emerging industries, but technology iterates quickly, so we must be prepared for long-term learning.”
Distinguish between "opportunities" and "traps"
Also in the AI chain, another group of people is being “harvested.” At the beginning of this year, BOSS Direct Recruitment disclosed a typical case of "recruitment-to-training": a recruiter used the pretense of recruiting "AI comic generation interns" and used words such as "As long as a college student can learn it quickly" to attract job seekers to submit their resumes, creating the illusion of "low threshold and quick money". After adding the job seeker's contact information, the other party further asked the job seeker to pay 2,000 yuan to "invest in shares" and claimed that he could become a full-time employee after completing 10 comics, with a monthly salary of up to 8,500 yuan. When the job seeker showed hesitation, the recruiter kept trying to induce him to pay by saying, "The income of a newbie's first drama is more than 100,000 yuan, and 2,000 yuan is already very cheap."
In fact, "recruitment-to-training" is not a recent scam. Instead, it has been constantly renovating its packaging with new concepts and new occupations. However, the core path has not changed: first attract the attention of job seekers through popular positions, and then in subsequent communications, in the name of training, cooperation projects, or certification, guide job seekers to pay. By the time the job seeker notices the anomaly, the so-called "admissions teacher" may have completely lost contact.
The experience of Li Ji (pseudonym) in Guangxi is more typical. In August this year, she saw a training recruitment post on a social platform. There were two classes to choose from, namely the AIGC comic drama full-process creation class and the AIGC comic drama OPC IP incubation class. When she couldn't afford enough tuition, the customer service immediately gave her a plan: "You can enroll in the school by paying a deposit of a few hundred yuan, and the rest will be paid back slowly in nine installments." Li Ji signed an installment contract, but later found out that it was a financial loan contract. The institution had already received all the tuition, and she had to repay the high-interest loan every month. "You have to pay back the class if you drop it, otherwise it will affect your credit score."
At the end of last year, job seeker Ren Min (pseudonym) came across an advertisement on social media, claiming to "learn AI from scratch, provide teaching and skills, and recommend employment." She gritted her teeth and paid 6,000 yuan, but the two-month training was extremely "watery" and the courses were almost all free tutorials that could be found online. The so-called "recommended employment" is even more of a lie. The tutor sent a few recruitment links but nothing happened. After she requested a refund, the instructor "packaged" her resume with his own work and urged her to go for an interview. However, once the other party asked more in-depth questions during the interview, Ren Min could not answer, and in the end he did not find a relevant job.
When "quick job" is packaged as a shortcut to a new career, and when "guaranteed employment" becomes a rhetoric for harvesting anxiety, those who are most eager to catch up with the trend are often the most likely to become targets of the gray industry chain.
Formal learning channels are actually not difficult to find. The reporter learned that Guangdong, Guangxi, Hunan, Jiangsu and other places have free AIGC training programs. There are also many schools and open universities that provide systematic study courses. Recently, Fang Zigen, a 71-year-old photography enthusiast and a student in the fourth phase of Shanghai Open University's AI short drama creation class, created an AI historical short drama at a low cost in just 15 days and was shortlisted for the third Hong Kong International Short Film Festival.
Luo Chong has his own methodology for "continuous learning". After work, he often exchanges creative experiences with colleagues and develops new tools. He said that it is best for AI creators to join a community of creators or enthusiasts, so that they can have various information about tools or algorithms.
However, under the tumult, distinguishing between "opportunities" and "traps" may be more important than learning to use any AI tool.
China Youth Daily·China Youth Daily reporter Chen Xiao
Source: China Youth Daily, October 13, 2026, Page 04
