Yuewen's "AI Virtual Human Lecture Hall": Good Stories Are the Future of the AI Era

Nowadays, the most popular content in the content industry is AI humanoid dramas.

It is a content form that has emerged from the AI comic drama track and is also a product of the leap - forward progress of large video - generation models.

As early as the second half of 2025, this content form began to show its potential. At that time, a group of practitioners believed that the ceiling of AI humanoid dramas was equivalent to that of the entire traditional film and television industry and higher than that of AI comic dramas. So, they decided to enter the market when many technical problems had not yet been solved.

Since January this year, the release of ByteDance's Seedance 2.0 and Shengshu Technology's Vidu Q3 has completely detonated this track. Problems such as the uncanny valley effect of character expressions, the difficulty in maintaining character consistency, and the incoherence of fight scenes have all been resolved with the progress of underlying large - model technologies.

Zheng Zhong, the person in charge of the Vidu solution, said that in the past year, large video - generation models have rapidly evolved from consumer - end applications such as "old - photo revival and transformation effects" to professional tools that can support the production of comic dramas and humanoid dramas. They have continuously improved in terms of consistency, aesthetics, and semantic understanding, "and have moved towards the field of serious creation."

However, the rapid change of technology has made the industry trend unpredictable. What kind of capabilities does a company need to remain competitive in this track? Will the currently popular AI humanoid dramas be the future?

In March, Yuewen Group held the "AI Humanoid Drama Lecture Hall". Following the "Comic Drama Lecture Hall" in November last year, Yuewen Group once again gathered front - line players in the industry to discuss the industry trends in the midst of change with the industry, academia, and the media.

Like every technological revolution, AI first brings cost reduction and efficiency improvement.

During the sharing, Meng Xiangyun, the founder of Aikan Interactive, mentioned that in late 2024, the team had a disagreement about the path: whether to prioritize content creation or tool development. Eventually, they concluded that "in the end, the competition in comic dramas lies in efficiency." In addition, in Meng Xiangyun's view, the core competitiveness of AI humanoid dramas is not just the improvement of picture quality, but the ability to stably output content in multi - location collaboration and large - scale production. To achieve this goal, tool - based and project - management capabilities have become key infrastructures.

Correspondingly, a voice in the industry believes that although the cost of AI humanoid dramas is higher than that of AI comic dramas, it is still more than ten times lower than that of traditional short dramas, and the cost per drama can be kept within 100,000 yuan.

This high - efficiency and low - cost feature has been passed on to the supply side, bringing about content changes and thus forming the advantages of AI humanoid dramas. In the past, due to budget and shooting conditions, large - scale scenes, science - fiction themes, and complex group scenes were often actively avoided. However, with the addition of AI, these contents have become important elements to attract users.

Meng Xiangyun said that when initiating a project, the team will first consider "spectacles", that is, visual scenes that were difficult to achieve in the past, and then develop the narrative around them.

When creation is no longer restricted by physical conditions, the competition will be pushed to a new dimension. Thus, content has once again returned to the field of vision of AI humanoid drama creation.

Huang Haonan, the founder of Jiangyou Culture, said at the meeting: "IP is like bullets. Without IP, there are no bullets."

In Huang Haonan's view, AI can significantly amplify production capacity, but it cannot infinitely generate high - quality content. The creative supply of creators has a limit, while the platform's demand for content continues to expand. Therefore, IP has become a key resource that determines the upper limit of supply.

In recent years, China's online literature, traditional film and television, and animation industries have accumulated a large amount of IP resources, but there is not much content that can be continuously developed. As AI lowers the production threshold, the development efficiency of these IPs has been significantly improved, and their value has also been magnified. IP is no longer just adaptation material but a long - term asset that can be continuously produced and expanded.

The in - depth development of an IP will also create a new competitive situation.

Ding Kuan, the general director of Jiuziyuan AI, believes that in the future, the core of content competition will shift from single works to "competition between IP universes."

In the AI environment, creators can continuously produce content around a worldview and build user stickiness through high - frequency updates, thus forming long - term value. His work "Gu Cheng Zhao" has achieved good data results on various platforms almost relying on natural traffic and has attracted the attention of overseas platforms, which he regards as an early verification of this model.

On this basis, the industry structure will also undergo new differentiation. On the one hand, AI tools have lowered the creation threshold, and small - scale teams have the ability to produce independently. Ding Kuan predicts that in the future, "small teams of 5 to 15 people" may become the main creative units. On the other hand, companies with IP reserves and large - scale production capabilities can still meet the larger - scale market demand through industrialized production. The connection point between the two is the continuous development and operation of IP.

As AI gradually popularizes production capabilities, the truly scarce resources are no longer "whether one can do it" but "what to do" and "what to do continuously." As the source and carrier of content, the value of IP has been further magnified.

Looking back at the present, is it still worth entering the field of AI humanoid dramas?

In terms of market size, AI humanoid content is in a stage of rapid expansion. At the same time, driven by both technological iteration and supply expansion, the participating guests generally believe that the time window for large - scale penetration of AI humanoid content is very close. Most judgments are concentrated within 3 to 6 months, with a maximum of no more than 9 months.

At the same time, the industry still faces several key challenges. Zheng Zhong mentioned that the model's understanding of the physical world still needs to be improved, such as the grasp of details like object weight and action mechanics, which will directly affect the sense of realism. In addition, issues such as copyright ownership and content supervision will gradually emerge during the rapid development of the industry.