A large number of AI-generated film and television works are about to be officially released to the public.

On September 14, a call for entries from the China Golden Rooster and Hundred Flowers Film Festival listed a set of materials different from traditional film submissions.

In addition to the finished film, poster and story synopsis, applicants are also required to submit the model toolchain, core prompts, AI generation process and engineering screenshots for technical verification.

The proportion of AI-generated footage in the submitted works shall not be less than 50%, and the feature-length works shall be no less than 60 minutes in duration.

These requirements come from the first independent AI film festival brand launched under the Golden Rooster system.

The Beijing International Film Festival and Shanghai International Film Festival have previously set up AIGC-related sections; this year, the Hundred Flowers Awards added an AIGC excellence promotion section for the first time, and the Golden Rooster further incorporated the duration of feature films, generation proportion and production records into the public collection rules.

On the other hand, changes have also begun to emerge in broadcasting channels.

Two weeks ago, *Post Journey to the West* was broadcast in the prime time slot of Hunan TV, pushing AIGC long-form content into this traditional TV broadcasting scenario for the first time.

On one hand, film festivals and award systems have started to set up separate access for AI works and establish verification rules; on the other hand, AIGC long-form content has begun to enter more traditional broadcasting channels.

AI film and television is moving from experiments in technology companies and creator communities to more formal industrial links.

To understand why this new production method has come this far, we must first return to the most traditional production site of a film: the film set.

After the Film Set Disappears

Film and television production is first and foremost a collective labor that relies on manpower, capital and physical space.

Actors must arrive on time, and the photography, lighting, art, costume and props teams need to collaborate simultaneously. A non-existent house has to be built, and a rain scene that only lasts for a few minutes requires the crew to mobilize water trucks, lighting equipment and dozens of staff members.

Once the machine is running, costs are incurred every day; if a scene is not finished, the subsequent plans will be delayed accordingly.

Liu Chi is a commercial director with more than ten years of industry experience, who has worked on advertisements, online films and the virtual human short drama *Liu Yexi*. She describes the changes brought by AIGC as "earth-shattering".

Earlier this year, the 45-minute AI animated film *Floating Life in Seven Days* completed by her six-person team won the Best Feature Film in the AIGC Film Section of the 16th Beijing International Film Festival. It took about one and a half months from script writing to post-production, and the core production of video generation and post-production took about three weeks.

In the theater project she is currently advancing, there are more than ten people in the AI generation link; including script writing, directing, storyboarding, concept design and post-production, the entire team still only needs 20 to 30 people.

"In the past, for a normal film and television process, a crew might need about a hundred people. Now with AI, it is enough to have less than ten people, no shooting is required at all, all processes are completed on the computer, and it is more like a post-production company."

The core of the change is not just the reduction in the number of crew members, but the mid-term reshaping of the film and television production process.

Director Chen Youxue describes the new process as a workflow between live-action film, animation and post-production.

Directors can use text instructions to mobilize the 3D model in Blender to generate white model dynamic storyboards, and confirm character positions, camera positions and camera movements in advance.

After the actors complete their real performances, their movements and expressions are transferred to the AI characters. Part of the work previously completed by set building, lighting and on-site scheduling is thus converted into digital asset production, shot decomposition and model control.

Theoretically, as long as the number of personnel and computing power are increased, the generation team can produce a large amount of materials simultaneously and compress the production process after the character assets are mature.

However, when actually making a film, Liu Chi still chooses to "polish scene by scene". The team needs to repeatedly check whether the characters drift, whether the emotions before and after can be connected, and whether the images serve the narrative, instead of generating a large number of random shots first and then looking for usable parts from them.

Machines improve the production capacity of images, but cannot replace aesthetic and narrative judgment.

This allows creators with backgrounds in fine arts, animation and concept design to enter AIGC film and television earlier.

Long Yitong, who is studying at the Central Academy of Fine Arts, began to experiment with AI images in 2023 and founded the Yuemo Studio.

In 2025, the AI feature film *Tai Sui*, an urban supernatural story co-produced by Yuemo, was launched on Tencent Video, which is also one of the earlier AI films in China to land on mainstream long video platforms.

Through this project, the team went through the full production process of AI feature films for the first time.

After that, they cooperated with a professional director team on the 30-minute film *Black Bird*, responsible for all AI content production. The film later won the Best Visual Creativity Award at the Aranya Theatre Festival.

In Long Yitong's view, after AI lowers the operation threshold, light, shadow, color, composition and artistic style have become more important instead. The new process requires a kind of "hybrid talent": who understands film and television, also understands fine arts, and can comprehend models, which is difficult to correspond one-to-one with the single type of work in the traditional film set.

Traditional companies have also begun to set up positions for this kind of composite work.

Zhang Zefan, a former film and television practitioner and teacher at a film and television college, told All-weather Tech that leading film and television companies such as Noon Sunshine and Youhug Media have begun to recruit "AI producer" positions.

In his view, this new type of work does not have clear boundaries at present. However, different from traditional producers who are in charge of funds, accounts, personnel and affairs, AI producers are closer to process producers in the animation industry: they decompose the creative team's intentions into shots and digital assets, coordinate generation, art and post-production links, and manage personnel productivity, computing power costs and final delivery quality at the same time.

The new production system has been in operation, and the industrial operation mode is also changing quietly.

Calculate the Overall Situation, Make Small Investments

Regarding cost reduction as the only value of AIGC is undoubtedly a crude view.

In the exchanges between multiple creators and All-weather Tech, "flexibility" appears more frequently than "low cost".

In the past, a film usually moved forward along the linear process of script development, creative team formation, shooting and post-production. Funds continue to be invested after shooting starts, and investors often have to wait until a large amount of costs have been incurred to judge from the sample film whether the characters, story and visual style are feasible.

This uncertainty is also one of the reasons why film and television investment has long relied on well-known IP, stars and mature director teams.

They cannot guarantee the success of the project, but can provide part of the recognizable credit for investors before the finished film is produced.

AIGC provides another way to place bets.

Chen Youxue will first let the team invest less computing power and energy to build the framework of a feature film, so that the creative team and investors can clearly see the characters, scenes, rhythm and story trend.

"In the past, we had to wait for the sample film to come out to see the effect of the film. Now a feature film can completely take shape within a month," Chen Youxue introduced to All-weather Tech. Based on this, directors can judge which paragraphs are worth strengthening, and then add actor performances or improve production specifications at fixed points.

He estimates that about 30% of the early generated materials may still be used in the subsequent finished film. The team can also take the Demo to meet investors, discuss whether to introduce better resource allocation, or stop the project before more funds are invested.

For investors, the same amount of early budget can test more projects at the same time, and then gradually screen according to the finished film prototype and platform feedback to leverage predictable and higher-quality projects.

Chen Youxue believes that under the investment framework of the film and television industry, this risk control is more important than infinitely reducing costs. Once a project shows commercial potential, the next step is no longer to consider how much more money can be saved, but what resources to add to amplify the returns.

The fact that decisions can be made in advance does not mean that the production process can be simplified all the way, and some links that determine the quality of the finished film still cannot be bypassed.

Liu Chi believes that script development cannot be accelerated infinitely by adding more generators. "A good script can be written for two years, ten years, or two months, but it cannot be produced in two days."

Post-production also does not disappear with the disappearance of the film set.

The links listed by Liu Chi include editing, color grading, dubbing, special effects, music and mixing, "all traditional post-production links remain". AI materials also need to deal with character drift, shot style and picture connection, which sometimes increases the revision work instead.

The fast turnaround feature of AIGC will also bring new commercial risks.

Commercial-level projects require repeated generation, screening and revision, and model credits are not negligible expenses. AIGC is indeed cheaper than large-scale live-action projects, but the investment for independent creators to complete a commercial work is still not low.

What is more tricky is technical depreciation. Chen Xing, founder of Beijing Tianjiao Star Culture Media Co., Ltd., told All-weather Tech that it is best for an AIGC project to be completed within two months from the date of establishment.

Once the cycle is extended, new models may require early materials to be remade, and production efficiency may suddenly increase by five times or even ten times.

"When the project is actually launched, people will feel that the picture quality is no longer amazing enough."

After the Decentralization of Production Rights: Decentralization and Invisible Threshold

The concentrated release of AI-participated project supply is approaching its eve.

Chen Youxue told All-weather Tech that several platforms he knows are producing mid-length dramas, long dramas and films, and many AI creators around him have also joined the projects, but the works have not yet been launched. He predicts that "by the first quarter of next year, there should be a number of excellent works coming out".

This is basically consistent with the signals released by iQiyi, Tencent Video, Youku and Mango TV.

In April this year, Sun Zhonghuai, Vice President of Tencent and Chairman of Tencent Online Video, revealed that Tencent is producing a dozen-episode drama and a 90-minute film with a full AI process.

At the end of August, *Post Journey to the West* was broadcast in the prime time slot of Hunan TV and Mango TV. The first 5 episodes of "Huaguoshan Chapter" were broadcast, and the remaining 25 episodes are planned to be broadcast in winter vacation, adopting the seasonal broadcast mode of simultaneous production, review and broadcasting.

Since June, Youku has launched the AIGC Online Story Film Incentive Plan, giving rewards based on the effective viewing performance after launch, with a maximum incentive of 1.35 million yuan for a single exclusive film.

*Erosion of Tide*, directed by Wang Ze, a graduate of the Vancouver Film School Shanghai, was shortlisted for the "Bao De Xi · iQiyi Film Theatre" this summer.

This is an attempt by iQiyi this year to further advance from AI short films to commercial feature films.

Compared with the previous "Bao De Xi · iQiyi AI Theatre" which mainly produced works of more than 15 minutes, the new round of plan directly raises the threshold to more than 60 minutes, and finally selects 10 AIGC Chinese commercial films for finished film production.

He told All-weather Tech that the project completed the final selection at the end of July, and is currently in the stage of script polishing, character design and asset production, with the ideal progress of completing production before the end of the year.

The development of feature films, which used to be calculated in years, is now emerging with new samples calculated in months.

After the threshold is lowered, the increase in supply is almost an inevitable result.

In March this year, Gong Yu, CEO of iQiyi, proposed that as AI reduces production costs and expands the number of creators, film and television platforms will shift from centralized content production to a more decentralized creator ecosystem.

iQiyi later established the AIGC Creative Center, expanding its cooperation targets to individuals, studios, production teams and university creation groups, and opening up directions such as theater films, online story films, long dramas and mid-length dramas.

The source of projects has thus been opened up. Individuals and small teams do not need to become mature film and television companies first, and they can also bring scripts, Demos and production plans to the platform.

iQiyi also opens part of its IP, computing power and investment and financing resources, and sets up different cooperation paths such as platform investment, copyright procurement and finished film revenue sharing.

However, the more costs the platform bears, the deeper its involvement in script modification, production progress and final delivery.

For example, the supplementary investment is funded by iQiyi for 10%-30%, and the review opinions are more for reference; the leading investment and full investment are respectively borne by the platform for 50%-60% and 100% of the funds, and creators need to refer to the platform's review opinions to advance production.

Long Yitong told All-weather Tech that she is not too worried about the competitive pressure brought by the rapid increase in AIGC supply. In her view, there are still not many teams that truly have stable creative capabilities and can continuously complete medium and long content.

She calls the mass-produced AI manhua dramas "AI Foxconn": texts enter the assembly line, characters and images are generated according to fixed templates, aesthetic and narrative judgments are handed over to the model as much as possible, and the team continuously expands production capacity by adding computing power and operators.

Once this workflow is implemented, low-quality supply will indeed expand rapidly. But high-quality long content can hardly copy this logic, "just like after the advent of cameras, everyone can press the shutter, but only a few people can become photographers."

This does not mean that the film and television industry has fully opened its doors to AIGC creators. Long Yitong has already felt this invisible threshold.

She quoted a judgment from a senior platform executive: "Most AI creators have not yet reached the professional standards of the film and television industry, and they have more benefited from a wave of technological dividends. Professional things should still be done by professional people."

In her view, if an independent AI studio does not have the endorsement of a traditional film and television team, the success rate of directly bringing a feature film project to the platform for review is not high.

Therefore, Yuemo did not package itself as an AI team that can independently handle feature films, but took the initiative to strengthen cooperation with professional directors and traditional film and television teams.

"We respect the professional barriers of traditional film and television, and give full play to our respective strengths."

Seize New Positions

Technological revolution is not just about changing a set of production tools, but also redistributing industrial positions.

In 2002, major Hollywood studios jointly promoted the establishment of the DCI Specification (Digital Cinema System Specification), which unified the file, projection and security standards for digital cinema.

Ten years later, Kodak filed for bankruptcy protection, becoming a symbol of the decline of the old technology chain.

At the same time, new industrial positions were vacated.

Companies such as Avid, Autodesk and DaVinci Resolve have gradually embedded themselves in the processes of editing, special effects, color grading and digital asset management.

They did not become film companies, but secured the standard position of "how films are produced".

This scene is being repeated in the