Image source: Generated by Unbounded AI

Ding Kuan and Long Xinyuan are both creators of AIGC film and television works. They use AIGC to create short dramas in China, and they are also among the earlier attempters.

Ding Kuan, who graduated from the Animation Department of Beijing Film Academy, created the first paid short drama "The Strange Things of Xing'anling" in China for more than three months. Within less than a week of online broadcast, it exceeded 35 million, and its revenue on the native side of Douyin exceeded 300,000, which led to his account gains of 80,000 fans.

After returning from studying in the United States, Long Xinyuan worked as an advertising director for seven or eight years, and later began to explore AIGC. After three or four months of self-study, the AI ​​MV produced by the Douyin platform gave more than 1 million likes. The AIGC paid short drama "35-year-old social animal apocalypse" recently produced by Long Xinyuan's team has been updated to episode six.

Ding Kuan and Long Xinyuan's current teams are within six people, both of which are small team operations, with division of labor including screenwriters, scripts, technicians, etc. This is also a team operation model that is more common for domestic AIGC creators.

In the view of Ding Kuan and Long Xinyuan, AIGC short dramas still face obvious pain points, involving many aspects such as talent training, business models, and technical barriers. There is still a lot of market space to be explored in content themes.

On the other hand, platforms such as Douyin have also begun some implementation support measures. In addition to some related traffic support, at the end of last year, Douyin and Jimeng AI and various universities jointly held the National College Student AIGC Short Drama Challenge Competition to encourage young blood to join generative AI creation, hoping to discover more AIGC talents.

At a recent AIGC short drama creation sharing session, Ding Kuan and Long Xinyuan's sharing aroused the curiosity of short drama practitioners. On and off the field, around the creation of AIGC short dramas, the creators not only raised some practical questions about how novices make profits and how to avoid pitfalls, but also put forward some professional discussions in practical aspects, such as how to overcome the technical defects in space mirroring in the process of using AIGC.

From the actual perspective, what has the development of domestic AIGC short dramas? What are their personal experiences and feelings? What are the costs and business models? What are their prospects for the future application of AIGC?

The development trajectory of AIGC short drama coincides with online movies. It is a regular basis to follow. Behind the creation of AIGC short drama "The Strange Things in Xing'anling" there is a clear creative strategy.

Ding Kuan took the 8-year development history of Chinese online films as a "model" and discovered that the development rules coincide with the trajectory of AIGC's creative content.

He divided the evolution of the online film industry into four stages: the first stage is centered on male frequency, fantasy, magic, and fantasy themes, and uses commercial gimmicks to open up the market; the second stage is to turn to more hardcore folk customs and monster types, representative works such as "The Five Immortals in the Northeast" and so on; the third stageThe paragraph is upgraded to a drama and a war film; the main theme is finally transformed.

In the current field of AIGC, just like the online movies and online dramas in 2014 and 2015, it also takes the theme of "Three Illusions" as the core. And with the support of AIGC technology, science fiction elements have been able to break through the bottleneck of early production.

Ding Kuan believes that the emergence of AIGC may mean that all past experiences will be repeated on a new carrier and platform. But above this, as a creator, you need to find a more breakthrough point of curiosity, which must be something that the audience has never seen before.

The Strange Things in Xing'anling set the story in the Northeast in 1962, using the Kwantung Army Fortress as the geographical coordinates, connecting the clues of the "uranium mine" left by the Japanese army with the history of national atomic bomb research and development.

The creative team deliberately retained 30% of the clear background explanations, leaving 70% of the narrative gap, and inspiring audience discussion by setting up controversial topics such as the "Five Immortals in the Northeast" folk legend and the alienation of Japanese zombies. This kind of "buried button" marketing design, on the one hand, avoids the creative stalemate of "anti-Japanese dramas", and also refers to the elemental innovation of "Japanese devil zombies" created by "The Dragon Search", ushering in a new breakthrough in the theme of tearing down the devils.

The work adopts a multi-line juxtaposition structure. On the one hand, it is a folk legend line, presenting the mysterious characteristics of the Northeast Horse Fairy culture; on the other hand, it is based on real historical materials, the "Battle of the Tiger Head Fortress" of the Soviet Red Army in 1945.

So Ding Kuan's team has planned several communication hotspots in marketing: details of the Kwantung Army Fortress, the five immortal system textual research that folk lovers are keen on, and the action scenes that the audience expects from zombie-type films, and dismantled these into independent short video communication units.

Ding Kuan emphasized that this creative model is essentially a four-in-one system integration of technological implementation, artistic expression, marketing communication and review norms, and seeking a balance between type innovation and market laws.

"From writing scripts, shooting, performing, editing, and then posting, I made nearly 200 films, and then this is very crazy. My two accounts are updated every day, and there are not 200 contents, and there must be 150."

In the frequency of daily updates, Ding Kuan found that AIGC video content became popular and needed to "get the hook". The underlying logic of algorithm promotion is actually a kind of thinking about human nature. Long Xinyuan also deeply agrees with this.

At the beginning of creation, they will reverse the trend from the marketing side, and at the same time, they will think about which hooks they should use to cause discussion on the audience's topic and which points the discussion on the audience's topic is concentrated.

For Long Xinyuan, from advertising director to AIGC short drama production, an important test was the AImv "Grandma Patriotic Version" he made, which was Xun on platforms such as Douyin.After the rapid fermentation, he firmly believed in the content advantages of self-media creation.

In his opinion, the reason why he was able to make the AIGC short drama "35-year-old social animal apocalypse" relies on the excellent compound talents of the team. If AIGC creation wants to complete a large-scale project, it requires combining the power of multiple super individuals, rather than aligning the "cheap labor" tactics.

"Because AI has a strong divergence, people who can control AI for high-quality creation must have strong comprehensive qualities and active thinking ability. High-quality AI creators are not cheap."

"AI can liberate productivity, but at present it is like a wild horse that has run away from its reins and needs to be controlled strongly

As an early AIGC practitioner from the fine arts major, graduated from the animation department of the Beijing Film Academy, and later transformed, Ding Kuan saw some current application status of AIGC in the early stages of the project.

On the one hand, film and television practitioners lack the awareness of AI technology, and AI engineers lack the artistic literacy of film and television, and there is a sense of separation of "technology does not understand art, art does not understand technology". It may be difficult to establish an effective dialogue mechanism in the two fields, so it is difficult to form a standardized technical process, and it is still in a exploratory stage.

"We are facing a 'three no' state in the market, and it can be said that there is no market environment, no data, no experience, and even no knowledge of what kind of subject story should be created. "

So, the current AIGC creation is still "workshop-style". The civil technology community is called "Juyutang". They are exploring fragmentedly, lacking data asset management and technical processes. For example, at present, the material style generated by AIGC is not unified, multi-modal content is difficult to coordinate, and cross-platform data flow loss is serious.

In terms of business model, the existing monetization path mainly focuses on two dimensions: advertising video production and knowledge payment training: the former is still subject to traditional advertising industry thinking, and AIGC is regarded as a low-cost alternative; the knowledge payment model often appears "19 The traffic monetization trap of 9 yuan limited-time courses.

The commercial closed loop of mature AIGC film and television content products is still far from going to go, and creators are still facing the risk of imbalance in input and output. Although AIGC is very strong, the current situation is still in the stage of crossing the river by feeling the stones.

"35-year-old social animal apocalypse"

In fact, AIGC currently has obvious pain points in talent technology, content production, technology implementation, etc.

As Ding Kuan mentioned, new creators with both artistic aesthetic ability and technical understanding are still scarce, and existing practitioners mostly show "polarization" characteristics: film and television backgroundTalents are still exploring the superficial technical applications, while technicians still lack the ability to understand lens language. Ding Kuan's team has a professional background in animation, but they have also received technical training step by step from scratch. In the current educational environment, systematic training methods have not yet been formed, and most practitioners are also exploring by self-study.

"In the future, AIGC talents, or AIGC directors, need to take into account both art and technology."

It is obvious that AIGC film and television creation faces some contradictions: first, the disconnection between typological exploration and market verification, past cases mostly stay in the technical demo stage, lacking reference for commercial success samples; second, historical data is insufficient, and effective creative decision-making rules cannot be constructed; third, the adaptation of the review mechanism and technical characteristics, and automatic content generation still has some potential risks in copyright ownership, value control, etc., and human resources are still needed to intervene in screening.

In addition, the audience engaged in the research on the model aesthetics of video generation large models raised a question, "AIGC dynamic mode has limited capabilities for space. But from the perspective of creators, film and television works have high requirements for control of mirroring, so how to use creativity to make up for technical defects?"

This is exactly the pain point of the technology implementation faced by AIGC. "Because of the limitations of the mode itself, only 20 to 30% of the mode can be completed in space mirroring, and the remaining 80% can only be done by local methods at present."

Also at the technical level, Long Xinyuan discovered that AIGC still has a relatively large bug in the image technology for the consistency of characters and the spatial position expression of scenes, which is also a major problem at present.

"No matter which model you use, you will feel a headache to make continuous narrative pictures. AI can liberate productivity, but at present it is like a wild horse that has broken away from its reins and needs to be controlled strongly."

In this regard, Long Xinyuan's response method is to complement each other through the advantages and disadvantages between different models. After using a large language model to perform literary pictures, he will perform secondary processing of the pictures, dealing with problems such as character consistency and bugs in spatial position scenes, involving the mutual combination of nearly ten AI models.

So, in Long Xinyuan's view, to make a work that can be accepted by the audience, it still requires strong control from the director and the team, which still consumes a lot of manpower and time costs.

Is the creation process of AIGC short dramas close to animation? Is the AIGC explosion approaching the critical point?

In Ding Kuan's view, the creation process of AIGC film and television is similar to animation, which is essentially animation in the film and television industry. In the film and television industry, the person who draws storyboards is not that high and the editor is very high, but in the animation industry, it is the opposite. The storyboards are usually drawn by the executive director or director himself, so animations often require very detailed dynamic rehearsals, which is a process and AIGC content productionThe process is very similar.

After he finished the storyboard script, he subdivided it into each team to do a short section, and finally bonded it to produce a complete AIGC work.

But the difference is that Ding Kuan believes that the overall atmosphere of the entire film and television industry is relatively impetuous, and it is unlikely to break a work in five years like dumplings. The AIGC creation process is not too complicated than the traditional animation industry. The retreat creation of small teams will be a more continuous and common model in the future. In a relatively short period of time, everyone will compete for efficiency.

This means that the launch and update of AIGC works will also be faster, and the competition may be more intense, presenting a linear competition.

Ding Kuan's next step is to make products in the context of Chinese locality, and he turned his attention to the female fantasy track.

Looking forward to the development of the AIGC short drama product form, Ding Kuan believes that the cost of making an AI short drama may not be lower than that of ordinary horizontal and vertical short dramas, but why does this product still develop? It is because when comparing with contents such as fantasy, magic, fantasy, science fiction, fairy tale, monsters, etc., it is found that it still has a great degree of cost reduction and efficiency improvement.

"Maybe it costs about one or two percent of the actual shooting, so the establishment of business logic will give birth to many works, and it will gradually form an aesthetic paradigm, which is the visual aesthetic system created by AI."

"AI aesthetics is a new aesthetic born after the two aesthetics of animation and live-action film and television. This aesthetic is what we will explore next." Ding Kuan said.

Long Xinyuan returned to Guangdong from Beijing in 2023. This year he began to come into contact with AIGC. By 2024, he felt a big explosion, "AI-generated videos became vivid." From the "landing period" to the present, Long Xinyuan has witnessed the changes in technology, from the cumbersome operational steps to the increasing simplification.

He felt that the entire AI was constantly sinking and would reach a state of full development in the future. This year he could see this kind of productivity liberation in the sinking market.

"As the person who made works earlier, I have provided a sample. With the pace of technology, when the indicators and links of each picture are getting fewer and fewer, allowing a person with ideas to open up creativity. This is the real AIGC explosion. I feel that I am increasingly stepping at this critical point."

Although the large-scale market has not yet been formed, in Long Xinyuan's view, AIGC will eventually usher in a period of "national creation", which not only means the maturity of the market, but also may mean that individual creativity will be magnified and presented.