Artificial intelligence is rapidly transforming China's short-drama industry, significantly boosting production efficiency and output. However, this surge is also causing concern among actors, voice artists, and even ordinary viewers, as the ethical and legal implications of AI's capabilities come under scrutiny.

In late March, several bloggers, including models and devotees of traditional Chinese attire, accused a popular AI-generated series, Peach Blossom Hairpin, of using technology to replicate facial features, outfits, and makeup without permission. These characters were widely distributed on short-video platforms, sparking a debate about the ethical use of AI in media production.

As a result of these complaints, the micro-drama platform Hongguo removed the show and banned its producer from uploading new content for 15 days due to insufficient proof of compliance with regulations on facial imagery. The decision underscored the growing tension between technological innovation and personal rights, highlighting the need for clearer guidelines and regulations, experts said.

This incident is not isolated. A number of celebrities, including actors Yi Yangqianxi and Yang Zi, have recently expressed their concerns, stating that they had never authorized the use of their images or voices in any AI-related dramas. They said AI-generated characters closely mimicking their appearance or voice constitute infringement.

Some cases have already reached the courts, with producers facing legal consequences for infringing upon personal rights.

Short dramas, or micro-dramas, typically shot in a vertical format with episodes lasting just a few minutes, have become immensely popular across China in recent years. By February, the short drama industry had 718 million monthly active users, with over 3,200 new short dramas released nationwide during the Spring Festival holiday alone, according to internet analytics firm QuestMobile. The explosive growth has been partly fueled by the convenience and accessibility of short-form content, which caters to the fast-paced lifestyles of modern audiences.

Generative AI is altering the production of these dramas. Industry analytics company DataEye reported that AI-powered animated shorts accounted for 38 percent of the top 100 comic-style dramas in January, up from 7 percent a year earlier.

The cost of producing a short drama has dropped from 50,000 yuan ($7,320) to 5,000 yuan per episode, and production time has been cut from two to four weeks to three to seven days. This dramatic reduction in cost and time has allowed more content creators to enter the industry.

However, the push for speed and cost reductions has led to mounting controversies. In the first quarter of this year, Hongguo removed 1,718 policy-violating animated short dramas, with 670 suspected of using people's facial information without consent.

These incidents highlight the challenges of balancing innovation with ethical considerations, as the technology outpaces the legal frameworks designed to regulate it. Legal experts said technological advancement should not come at the expense of personal rights.

More frequent

Zhao Qi, a judge at the Beijing Internet Court, said disputes over AI-related infringements of portrait and voice rights are becoming more frequent as the technology evolves.

Some defendants claimed they had used AI face-swapping technology just for fun, not profit.

But Zhao said that publicly broadcasting someone else's facial features without permission is prohibited under the Civil Code. "The distinction between personal use and public distribution is crucial, as the latter can have significant implications for the individuals involved," she said.

On March 23, the court disclosed details of a notable case in which Zhao ruled in favor of an actress whose images were violated by two companies that used AI face-swapping tech in a short drama. The court ordered the defendants to apologize on their video platforms and pay compensation to the actress.

The dispute began when the actress found her likeness mapped onto a character in a 90-minute drama, misleading audiences into thinking she had performed in it, and sparking online discussion. The series also appeared on another company's platform, leading her to sue both platforms for unauthorized, profit-driven use of her image.

Zhao said the production company argued the AI-generated character coincidentally resembled her. It claimed it only used prompts like "pretty journalist" without directly referring to her, and that the resemblance was "statistically probable". The production company said it quickly removed the segment after her complaint, causing no harm.

The streaming company argued that it legally obtained distribution rights and was not responsible for content creation.

However, when asked to demonstrate their AI production processes during the trial, the companies failed to do so due to technical and account issues, Zhao said.

"Even if an AI-generated face differs slightly from the original, it's still infringement if the public can recognize the person and no permission was given," the judge said.

"The streaming platform, despite having distribution rights, was also held accountable for failing to review content and prevent infringement."

Similar disputes have been ended through negotiation.

In March, for instance, an AI-generated fantasy short drama with over 23 million views sparked controversy for allegedly misappropriating a model's likeness and makeup style designed by a photographer. After the model, surnamed Ma, and the photographer filed complaints with the streaming platform, the series was taken down. Following negotiations, the two sides reached a settlement regarding compensation for the infringement of portrait rights and copyright.

Growing concerns

Only a year ago, AI-generated videos were rough, choppy, and visibly artificial, said Chen Peng, a short-video industry professional. They mainly consisted of single images or clips lasting just a few seconds, according to reports.

"But today, the technology has been completely transformed. It has leapfrogged in both the efficiency and quality of image and video generation," Chen said.

"Powered by next-generation large generative models, it can now produce complete short dramas with visuals, scenes, background music, and voice overs," he said.

"Thanks to breakthroughs in character consistency, long-text understanding, and fine-tuning by creators, the image quality, motion, and details are now virtually indistinguishable from real footage," he added.

The rapid development of technology has also lowered the barriers to becoming a creator and reduced production costs.

Zheng Ning, head of the law department at the Communication University of China's Cultural Industries Management School, said AI-generated short dramas can now achieve full-process generative AI coverage — from scriptwriting and character generation to scene rendering, voice overs and music production, editing and promotional distribution.

"This comprehensive integration of AI into the creative process is reshaping the industry making it more accessible, but also more complex," she said.

According to her, this compresses traditional film and television production timelines from months or even years to days or hours. With the barriers to entry removed, amateurs and tool providers — non-professionals — can also get involved in creation.

Zheng also pointed out that traditional film and television strictly follow a "review before release" principle, requiring creators and producers to go through a full process of content filing, review, and access approval, keeping the distribution scope and release cadence under control.

"But with AI-generated short dramas, the cumbersome review process has been simplified, enabling dozens of episodes to be updated in a single day, and hundreds of pieces of content to be generated in bulk, spreading extremely fast in cyberspace," she said.

The weak legal awareness of AI users has led to infringements of portrait and voice rights, Zheng said. Internet platforms, faced with the overwhelming output of AI-generated short dramas, have also failed to fulfill their responsibility to review content, while their review technologies still have significant shortcomings.

"Many platforms often rely on a reactive approach — removing content only after receiving complaints — rather than proactively identifying infringements. As a result, by the time an infringement is addressed, the short drama has already gained massive viewership and generated substantial revenue," she added.

Way forward

Zheng and Zhao both believe comprehensive governance of the short drama industry is needed, as frequent infringement issues not only violate creators' rights but also hinder the sector's long-term development.

"AI users should strengthen their legal awareness," Zheng said, advising them to conduct self-reviews to ensure their content does not infringe others' rights, and to obtain authorization before use.

"Platforms also need to prioritize preventive measures and provide easier reporting channels for users to protect their rights," she said. This could be achieved, she suggested, by enhancing the review process and applying cutting-edge technology to block infringing content more accurately and efficiently.

Judge Zhao said platforms cannot shirk their responsibilities by using the high volume of uploads as an excuse. "Technological advancements have raised the legal duty of care expected of platforms as content distribution hubs," she said, noting that the increased responsibility requires platforms to adopt more sophisticated tools and strategies to manage content effectively.

Both Zheng and Zhao support platforms establishing more convenient complaint reporting systems, and taking stronger actions against users who repeatedly post infringing content.

Zheng proposed a blacklist for infringing parties and an industry ban on repeat violators.

She also emphasized stronger enforcement of regulations that require AI-gene-rated content to be labeled and filed on record before release. Industry associations should also strengthen self-regulation, while technologies to detect voice imitation and face-swapping should be continuously upgraded, she added.

"AI makes it easy to produce a short drama that 'looks good', but harder to create work that 'truly moves people'. This urges all creators to resist the inertia of algorithms and remain committed to observing humanity," Zheng said.

"The focus on the way of promoting the industrial development should shift from 'using AI to replace creation' to 'using AI to empower creation'," she said.

By prioritizing human-centric storytelling, creators can harness AI's potential while preserving the artistic integrity of their work, she added.

Zhao said the courts will continue to balance technological innovation with the protection of personality rights, ensuring neither is compromised. This delicate balance will be crucial in navigating the future of AI in the creative industries, as both innovation and ethical considerations continue to evolve.