China's Entertainment Sector Surges as AI Dramas Go Mainstream

China's entertainment sector observes a notable shift as artificial intelligence-generated video dramas gain increasing prominence. A 30-episode AI-produced adaptation of the classic Chinese novel *Journey to the West* recently premiered on Hunan Satellite Television, a state-owned broadcaster, and its streaming platform Mango TV. This marks the format's graduation to prime time, following its extensive presence across social media.

Credit: UNSPLASH

The production of this series, with each instalment lasting approximately 40 minutes, progressed from commission to broadcast in a mere three months. The director has indicated that later episodes will undergo continuous refinement based on audience feedback. This rapid development capability highlights a significant change within the industry's production methodologies, pointing to a new era for content creation.

The premiere impacted China's media and entertainment sector, which had experienced a period of stagnation. Shares in Mango Excellent Media, the Hunan broadcaster's listed entity, reached their 20 per cent daily trading limit twice before stabilising. This surge added over USD 1.9 billion to its market value, indicating renewed investor interest. Guo Guo, a Beijing-based producer, stated that this event sends a clear signal of policy support, confirming AI drama's entry into the mainstream.

The widespread adoption of AI-generated content (AIGC) has accelerated across the content industry. Guo, now working on AIGC projects, detailed that a team of 10 can complete a 100-minute vertical drama within a week. More complex, long-form productions can achieve a rate of one 45-minute episode per week.

According to data from the China Netcasting Services Association, 367,000 microdramas were released on Chinese platforms during the first half of the current year. More than 74 per cent of these productions were AI-generated, equating to an average of over 1,500 new AI dramas daily. Production teams are actively adjusting to this altered landscape.

Conventional production elements, such as art departments, set crews, cinematographers, lighting technicians, and actors, are no longer necessary for fully AI-generated projects. This eliminates a substantial portion of traditional overheads, significantly reducing costs. Liu Zhanhao, another producer now focused on AI content, noted these cost reductions as "exponential."

The roles within these new production models now include "card pullers." These individuals function as part prompt engineers and part visual editors, feeding reference images and commands into AI models. They repeatedly adjust these inputs until a suitable shot is generated. Liu's next project, a 120-minute science-fiction mystery, classified as a premium AI drama, is slated for release on Mango TV this autumn. He observed that large model developers, or Big Tech devs, are now claiming a share of film and television industry profits.

Liu's team adapted to the technology quickly, moving from ByteDance's Seedance 2.0 to Seedance 2.5 in July as soon as it became available. Most of the budget for their projects is allocated to the generation process itself. However, video generators still face challenges. Maintaining character and setting consistency, interpreting intricate narratives, and overcoming limitations on clip length remain obstacles. Guo shared a similar observation, noting that AI models still struggle with scenes requiring high "emotional density."

And platforms are actively encouraging more AI-generated productions. Streaming services, including Mango TV, iQiyi, Tencent Video, Douyin, and ByteDance's Hongguo platform, provide cash subsidies and algorithmic traffic boosts for AIGC projects. Volume of content has increased more rapidly than its quality.

Liu stated that audiences exhibit surprising tolerance for production "glitches," often watching what he termed "brain-off gratification dramas." These viewers seek plot twists, melodrama, and exaggeration, prioritising an immediate emotional response. Ivy Yang, founder of Wavelet Strategy, a consultancy, suggested that short dramas are particularly susceptible to becoming formulaic AI output.

Yang noted that, even before AI, production logic favoured templates rewarded by algorithms, such as stories involving domineering CEOs, revenge arcs, humiliation followed by comebacks, or time travel. These were produced at the lowest possible cost, highest speed, and with the aim of delivering a strong dopamine hit. She added that AI has driven the cost of replicating these formulas down to near zero.

This development, Yang contended, results in a systemic increase in tolerance for low-quality content. She also expressed concern for younger viewers, who are still developing their narrative and aesthetic judgement. The shift, she argued, erodes patience for ambiguity, set-up, or any content requiring viewer engagement beyond passive consumption.

Producer Liu anticipates that inferior AI productions will be eliminated as competition intensifies, much like copycat mobile phones in China eventually disappeared. The country's cyberspace watchdog reported this week that it removed millions of pieces of AI content, including fake news, explicit material, violent content, impersonation, and safety violations involving minors, as part of a campaign against generative technology misuse.

Bryce Gu, a producer of live-action series, described the film and television industry as "anxious" regarding AI. He observed that distributors now demand AI involvement in nearly every project, even for those where the technology is unsuitable. Gu noted that formidable barriers to entry previously protected the industry, but AI has significantly lowered them. While AI can accelerate recognition for talented individuals, he cautioned that it cannot create talent itself.

• AI-generated dramas are entering mainstream Chinese television, exemplified by a 30-episode series produced in three months.

• The shift to AI production significantly reduces costs by eliminating traditional film crew roles.

• Despite rapid production and investor interest, concerns remain about content quality and its effect on audience expectations.

• Chinese platforms are incentivising AI-generated content through subsidies and traffic boosts.

• Regulators are addressing misuse of AI content, while industry professionals express both anxiety and cautious optimism about the technology's long-term impact on talent.

Source: SCMP