Woofun AI reports that the Chinese short-form drama industry has undergone a fundamental structural shift, transitioning from a paid-content model to an advertising-driven ecosystem before rapidly pivoting toward artificial intelligence generation. In the first quarter of 2026, the China Internet Audio-Visual Association released data indicating that approximately 128,000 short-form dramas were launched across the nation, a figure that underscores the sheer volume of content flooding the market. Of these 128,000 titles, 122,000 were generated by AI, representing a staggering 95% share of total production capacity. This dominance marks a decisive moment where AI-generated content has become the primary engine of the industry, relegating traditional real-person productions to a supporting role on mainstream platforms. The trajectory of this sector reveals a two-stage evolution: first, a commercial logic shift from direct user payments to traffic monetization, and second, a technological displacement where algorithmic creation supersedes human performance. The initial phase saw short-form dramas abandon the 'top-up to unlock' revenue model, which attempted to replicate online literature monetization, in favor of a 'free content plus advertising' strategy. This transition was driven by the attention economy of short-video platforms, where algorithms prioritize content that generates clicks, engagement, and sharing over high-barrier paid offerings. Consequently, the metric for content value shifted from individual willingness to pay to the ability to attract mass engagement, forcing creators to accelerate plot pacing, increase twist frequency, and amplify character conflicts to compete for limited user attention. This competitive environment fostered a consensus among creators to capture viewers within the first 3 seconds, deliver a plot twist within 30 seconds, and retain attention for at least 1 minute. To meet these demands, storylines increasingly focused on social mobility, revenge, extreme conflicts, and stark contrasts, elements designed to trigger immediate emotional responses. Simultaneously, local governments across China, from Hengdian to Zhengzhou, Xi'an to Chengdu, established production bases and industrial parks to leverage the short-form drama industry as a driver for cultural and economic growth. These initiatives aimed to stimulate film and television production, venue rentals, costume and makeup services, post-production, and job markets, creating a robust ecosystem centered around real-person filming.

However, the rapid ascent of AI technology has disrupted this localized industrial structure, introducing a second major change in the underlying logic of the industry.

Woofun AI data shows that the number of new real-person filmed dramas launched in the first quarter of 2026 decreased by approximately 75% compared to the previous year, signaling a severe contraction in traditional production. This decline is not isolated to a single region; filming bases in Hengdian, Zhengzhou, and Xi'an are reporting a significant reduction in production teams and empty studios. Many small and medium-sized teams have begun shifting their focus to producing AI comic dramas and AI realistic-person short-form dramas in large quantities to survive the market shift. The emergence of AI-driven short-form dramas has transformed aspects of content production from a labor-intensive film and television industry into a software-based industry, fundamentally altering the local industrial structures that were built around real-person filming. AI-generated short-form dramas are currently categorized into two distinct types: AI comic dramas, which feature anime or manga styles suitable for exaggerated fantasies and unique worldviews, and AI realistic-person dramas, which utilize AIGC technology to generate characters with highly realistic appearances, lighting, and movements. The dominance of AI comic dramas is not necessarily due to an inherent superiority of the anime style for short-form content but rather reflects the current technical limitations of AI in generating realistic human performances. While AI has achieved considerable success in creating static images, it still faces significant challenges in ensuring the realism of human performances, the coherence of actions, the consistency of timing, the logic of lighting and shadows, the synchronization of lip movements, and the expression of complex emotions. When cameras capture close-ups, subtle expressions, or interactions among multiple characters, AI-generated content often suffers from stiff movements, distorted expressions, and the 'valley of fear' effect, failing to meet audience expectations for real-person film and television works. Any deviation in quality can quickly undermine the immersive experience, making AI comic dramas a more viable option as their stylistic approach accommodates these technical shortcomings. If AI technology advances to consistently generate high-quality realistic-person performances, the industry landscape may undergo another round of changes, but for now, the competition between AI comic dramas and real-person dramas is not entirely homogeneous. The rapid rise of AI comic dramas is largely attributed to their ability to leverage lower production costs and faster update rates, characteristics that align perfectly with a content market reliant on algorithmic recommendations and traffic competition. In an environment where content supply increases indefinitely, the ability to produce content becomes less valuable than the ability to attract attention. The transition from a paid-to-view model to an advertising-based model laid the groundwork for this explosion of AI-driven content by prioritizing volume and engagement over individual transaction value. This reshuffle has inevitably brought fluctuations and panic, exemplified by the 'AI artist database' controversy sparked by iQiyi in April 2026. On April 20, during the '2026 iQiyi World Conference,' the platform announced the launch of an AI artist database, stating that more than 100 artists had joined. CEO Gong Yu mentioned in his speech that while live-action performances would not disappear, he questioned whether 'some completely 100% real and physical works might one day be designated as World Cultural Heritage or intangible cultural heritage.' He further noted that in traditional film and television production, actors often work for several months at a time with daily hours exceeding ten, leaving little personal space, whereas AI technology could increase an actor's annual output from four to fourteen works while allowing them more rest. Following the announcement, iQiyi revealed that hundreds of celebrities had joined the AI artist database on its 'Nadou Pro' platform.

However, as news spread on social media, the hashtag 'iQiyi is going crazy' quickly topped the Weibo trending list, and many relevant actors posted clarifications stating they had not signed any agreements related to AI. iQiyi subsequently explained on Weibo that joining the 'Nadou Pro' platform simply meant expressing an interest in cooperation. This incident highlights the difficulties faced by the long-video industry regarding cost structures, as platforms losing money cannot afford to ignore the potential of AI to reduce costs and increase efficiency. Some industry observers argue that using celebrity IP to attract key resources, a strategy that worked in the traditional entertainment industry, is no longer effective in the current landscape. Gong Yu's comments suggest that with AI, production teams no longer need to worry about scheduling conflicts or exorbitant actor salaries, reflecting the deep-seated cost pressures in the long-video sector. Micro-short-form dramas, as a type of short-video content, have exerted a more severe impact on long-video content than many realize. Over the past decade, the long-video industry operated on the assumption that users were willing to spend significant time watching long-form content and pay for high-quality content through membership subscriptions.

However, the rise of short-video platforms has increasingly challenged this assumption, as a large amount of content consumption now occurs on short-video platforms through clips, plot summaries, and highlights. The value of long-video content is increasingly realized and consumed on short-video platforms, with a significant proportion of views and popularity coming from secondary creations and promotions. User behavior has also undergone profound changes; while long-video platforms once enjoyed stable engagement and high loyalty, the growth potential of daily active users and monthly active users on mobile devices has significantly narrowed, with some platforms facing contraction. Total user engagement time has not increased indefinitely but has been redistributed among short videos, live broadcasts, games, and social media. In this context, it is becoming increasingly difficult for long-video platforms to attract new users or maintain the viewing time of existing users. Compounding this challenge is the fact that the cost structure of the long-video industry has hardly changed. Producing a high-profile TV drama often requires hundreds of millions of yuan in investment and takes several years, while large-scale variety shows require continuous investment in venues, guests, and production costs. Purchasing copyright also incurs substantial expenses, and these costs must be incurred in advance regardless of user numbers. Membership subscriptions, once considered the most stable revenue source, are no longer as reliable as users' willingness to pay tends to saturate and price increase room becomes limited. Long-video content also faces uncertainties, such as delayed releases or postponements, which significantly diminish value after three months. Long-video platforms have thus fallen into a typical operational dilemma characterized by high fixed costs and limited growth potential. In the past, scale expansion helped spread these costs, but as user growth slows, high production costs begin to erode profit margins, resulting in poor ROI even for highly popular dramas. Gong Yu's push for AI aims to address this business model, desperately needing AI to attract capital investment, but his approach inadvertently provoked public backlash. The AI artist database has triggered strong resentment and resistance among the public towards AI-powered actors, with posts on Weibo expressing 'physiological aversion' towards AI-generated faces. One reason for this is the excessive production of AI-generated dramas, leading to widespread instances of 'face duplication' and 'face theft.' The industry has even coined a terrifying term, 'piecing together faces,' referring to the use of faces from multiple real-person actors in the same drama. The purchase of ordinary people's portrait rights for AI short-form dramas has given rise to a new gray market, forming a complete 'face collection' chain. Beyond script copying and music piracy, unauthorized face replacement is a common issue in AI-driven dramas. Currently, protagonists of AI short-form dramas are mostly based on standard faces generated by large-scale video models because these faces are less likely to resemble real celebrities.

However, since there are not many copyrighted face samples in these models, the faces of characters in short-form dramas often appear very similar. Although AI comic dramas have overtaken real-person dramas, they still rely on real people's faces, placing performance at the forefront of these dramas. Traditional actors are a medium that relies heavily on the physical body, where performance is a combination of dialogue, expressions, time, context, and randomness. The same scene can produce subtle but crucial differences under different circumstances, and mistakes, pauses, and emotional fluctuations are part of the authentic quality of performance. The reason performance is moving is not its perfect reproduction but its uncontrollable and non-repetitive nature. The emergence of AI-powered actors has transformed performance from a physical medium into a digital one, where acting skills are merely the result of parameter optimization and characters become model assets that can be called upon, assembled, and reproduced. In this transformation, performance has lost its status as an 'event' and has become a standardized production process. Audience perception of actors' performances involves several layers, starting with authenticity. Currently, AI technology cannot create completely realistic performances, but if it becomes advanced enough that human senses cannot distinguish between real and artificial performances, people may not see a difference. We cannot oppose AI-powered actors solely on grounds of perceived authenticity.

Furthermore, emotional reactions such as laughing, crying, or tension can be replicated if AI-powered actors convey emotions that resonate with audiences. We invest emotion in watching fictional anime characters, and our empathy mechanisms do not stop functioning just because characters are not real. From an immediate-perception perspective, emotional reactions to AI-powered actors may be very similar to those for real actors.

However, deeper analysis reveals subtle differences. Human perception can be divided into primary and secondary levels; realizing a performance is entirely virtual may create a sense of intellectual distance or add aesthetic and philosophical dimensions to the emotional response. The presence of a performer's physical body embodies effort, risk, vulnerability, and humanity itself, irreplicable aspects that influence emotional reactions. When the physical body is absent, a subtle sense of absence may persist, and even if emotions are accurately triggered by technology, a direct connection to humanity may be lacking. The premise of this perception is the physical body itself; you are a physical being, and your identification with actors was based on their physical presence, dedication, repeated rehearsals, and emotional investment. In the movie 'The Black Swan,' the lead female dancer played by Natalie Portman invested tremendous effort and endured rigorous training to become a perfect dancer, ultimately suffering a mental breakdown. Watching such a performance moves us deeply because it touches us emotionally, reflecting the complexity of human emotions where hope exists amidst fear and jealousy is inherent in love. It is uncertain to what extent AI algorithms can perfectly capture this complexity in various emotional expressions. Audience preferences are changing, with more people valuing appearance and traffic rather than acting skills. It is hard to say whether audiences valuing these factors will accept AI-powered actors without hesitation, or whether the next generation born in an AI-dominated world will naturally accept them. While it is impossible to predict whether AI-powered actors will be widely accepted, one thing is certain: if they are, it will indicate that human aesthetic abilities and emotional richness are declining. Strictly speaking, all AI-generated content lacks copyright, as it is impossible to trace the identity of the creator. This is a global problem, as the copyright system is essentially a product of the industrial and print eras. Although adjusted for electronic media, a new wave of challenges is emerging, leaving many aspects of the cultural industry in a state of turmoil. When watching theater, we see the stage in person; when watching movies, we see recorded footage.

However, a performance is much more than what we see; it requires the joint efforts of screenwriters, directors, actors, photographers, producers, and others. What we see is the final result of this collaborative process. We are amazed by Natalie Portman's outstanding performance, but we cannot see all the behind-the-scenes work. The behind-the-scenes work includes the script, production, and other structural elements, which are stable and predictable.

However, performance itself is dynamic, unpredictable, and fleeting. Our current debate about AI-powered actors focuses only on the front end of this process, but what is more critical are the underlying text and production elements. By text and production, I mean the three different systems that constitute a theater performance: the script, production, and performance itself. Compared to the unpredictability of performance, these two elements are relatively stable.

However, in times of cultural upheaval, the traditional text and production systems are disrupted, and new systems and symbolic frameworks must be established. In terms of text, AI-driven short-form dramas attempt to automate narrative production. AI-driven short-form dramas can be seen as the first large-scale commercial experiment of generative AI in the cultural industry. What they produce is not just text, images, or videos but stories, one of the oldest forms of cultural content. In the past, narrative was always considered an activity that relied heavily on human creativity. From myths and epics to novels and movies, the conception of stories, the creation of characters, and the development of plots have all been seen as products of human imagination and accumulated experience.

However, the automation of narrative production by AI challenges this long-held view, suggesting a future where the very fabric of storytelling is redefined by algorithmic logic rather than human intent. The implications of this shift extend beyond mere production efficiency, touching upon the fundamental nature of cultural creation and the role of human agency in the arts. As the industry continues to grapple with these changes, the tension between the efficiency of AI and the authenticity of human performance will likely remain a central theme in the evolution of the cultural landscape.