Here's the English translation:
Side A: Producers make a drama in 7 days, no time to indulge in AI's big promises
On top of the industrialized production capabilities developed over decades in web novels and film/TV, short drama production has reached maturity within five years. In the eyes of many producers, there's no room left for AI in the tightly integrated short drama production chain.
Looking at the script stage, the underlying logic of short drama scripts is formulaic, which should be the comfort zone for large language models. Indeed, many AI writing products have appeared on the market, such as Wawa Writing and WriteWise.
However, in the entire short drama business model, the content is actually the least important part. The production logic of short drama companies has always been to buy enough scripts, shoot enough dramas, and bet on the probability of a hit emerging.
In terms of content quantity, short drama platforms like Jiuzhou already have their own writing teams and editing teams, able to stockpile short dramas at extremely low costs. The total production cost of a short drama is between 300,000-1 million yuan, with major platforms typically paying 10,000-30,000 yuan per script, while smaller platforms may pay only a few thousand yuan. AI has no price advantage and instead increases learning costs.
Scripts are not scarce; good scripts are scarce, but good scripts are hard to define. The reason short dramas have formed a probability-betting mechanism is precisely because under the algorithmic mechanisms of Douyin and Kuaishou, there is no universal formula for creating hit content.
Rather than delving into the elusive user psychology and algorithmic black boxes, a more practical approach for short drama producers is to find a screenwriter who frequently produces hits, betting on higher probabilities, or to quickly copy, shoot, and distribute content immediately after a film, TV series, or short drama in a particular genre becomes popular, directly imitating the recently validated path to success.
For example, when director Jia Ling's movie "YOLO" was a hit during this year's Spring Festival, short drama companies quickly launched "YOLO: Glamorous Transformation". There are also examples of riding on the coattails of popular titles, such as "The Marriage that Rides the Wind and Breaks the Waves" piggybacking on "Sisters Who Make Waves", and "With Feng Xing: Battle Goddess Mom Drops the Act" playing off "With Feng Xing", and so on.
To bet on probabilities, short drama production capacity must be expanded. In the filming stage, speed is paramount, and in this aspect, current AI technology lags far behind humans.
A clear contrast is this: Xianren Yikun's "Mirror of Mountains and Seas" took half a year to produce 15 minutes; while the filming cycle of a Hengdian short drama has been compressed from 2 weeks to 7 days, or even 3 days or 1 day. The crew members may not even have time to remember each other before the shoot is completed and the crew disbands. They have no need or desire to spend time exploring how to replace traditional filming with AI.
Even when AI face-swapping short dramas that went viral overseas crushed native overseas short dramas in terms of cost and timeliness, they quickly faded into obscurity due to lack of user acceptance. This is another characteristic of the short drama industry - following the money.
Compared to the production cost of nearly $300,000 for shooting with overseas actors, the entire production process using AI face-swapping only costs tens of thousands of dollars, reducing costs by 5 times. But the final effect couldn't withstand market validation. Due to numerous microexpression flaws in AI face-swapping technology, the user retention and payment rates for face-swapped short dramas were very low. The "2024 Short Drama Overseas Marketing White Paper" released by TikTok for Business mentions that in terms of content, local original dramas and translated dramas coexist, but hit dramas are mainly concentrated in local original productions.
This also means that although short dramas have different acting requirements compared to TV series and movies, allowing for roughness and absurdity, they need to be sufficiently direct, ensuring fast pace, emotionality, and high information density. However, what current AI video is best at is empty shots and slow motion.
Character performance is also a major pain point in current AI production. In the process of exploring AIGC works, Xianren Yikun discovered that "through text-to-image and image-to-video methods, AI videos can now control action consistency and scene consistency, but action amplitude and character performance still need improvement." Users often watch revenge-themed short dramas to project themselves into the story; AI-flavored fake characters would only break immersion.
To avoid character performance issues, recently launched AI short dramas like "Mirror of Mountains and Seas", "Sanxingdui: Future Revelation", "Fantasy Store", etc., all feature non-realistic fantasy or sci-fi themes, emphasizing imagination and visual impact, which are suitable for AI generation and display.
Looking at each link in the entire production chain, under the high-speed operation mechanism in Hengdian, short drama producers pursuing quick, cheap, and easy production simply don't have time for AI.
Side B: AI editing and traffic buying, more than 10 times faster than humans
While short drama producers show little interest in AI, those distributing short dramas have begun to enter the AI exploration stage.
After all, for short dramas, traffic determines life or death. In paid and CPS short dramas, the amount of traffic consumed basically reflects the revenue of the short drama. So short drama success reports, besides announcing their recharge amounts and revenue share amounts, more commonly use consumed traffic as a statistical dimension, i.e., how much money was spent on advertising and buying traffic.
This is because during the traffic buying process, as long as the preset ROI is reached, buyers will continue to purchase traffic, pushing for higher payments. Roughly calculated, taking an ROI of 1.2 as an example, this means that for a short drama to receive 120 yuan in user payments, it needs to invest 100 yuan in traffic acquisition costs. If a drama's total consumption is less than 1 million yuan, it means the revenue is less than 1.2 million yuan, the traffic buying profit is less than 200,000 yuan, not even covering the production costs.
However, as the highest-cost segment in the short drama industry, traffic buying for short dramas is also currently the industry's biggest pain point.
From a macro data perspective, as short drama producers increase and the overall short drama traffic growth slows down, the price of buying traffic is getting higher and paid viewership is becoming harder to obtain.
Data shows that the overall scale of traffic buying has begun to shrink.