What’s happening?

In recent months, a wave known as “microdramas” has spread: short series, with episodes no longer than two minutes, designed to be watched vertically on phones.

The new development is that many of them are no longer filmed at all. AI creates them from start to finish.

China is leading the way. In the first quarter of 2026, around 128,000 short series were released, more than 95% of them AI-generated.

Among them was a short film titled “Mother’s Monster.” Its production company, Zitomate, says it took less than 72 hours to make and cost no more than $2,000. Reposted clips from the film have racked up nearly one billion views.

The wave has moved beyond phones, too. On August 31, the series “The Later Journey to the West” premiered in China—the first fully AI-generated long-form television production. Shares in its producer, Mango Excellent Media, rose 64%, according to Bloomberg.

How are they made?

The idea is simple in principle. The creator writes the story and describes the scenes, then AI models generate the characters, movement, audio, and backgrounds.

One of the best-known tools is Seedance 2.0, from ByteDance, TikTok’s parent company. Launched in February 2026, it exceeded market expectations for realism.

Cost is what has turned the industry on its head. A minute costs around $30 to produce, meaning a 50-minute show can be made for about $1,500 in a week, compared with a full month and an entire crew for traditional production.

Revenue comes from the usual sources: advertising, pay-to-unlock episodes, subscriptions, or selling distribution rights to platforms.

Why now?

Three reasons.

First, AI models have advanced at remarkable speed. A year ago, generated clips were easy to spot; today, they are harder to distinguish.

Second, platforms need an endless stream of content. Phone algorithms consume content at a pace human production cannot match.

Third, there is government support. Several Chinese cities are competing to attract startups in the field, offering them funding, computing power, and specialized centers.

Abundance doesn’t mean success

Here we come to the most important figure in the whole story.

Of the 221,900 AI-generated productions, only 1,055 surpassed 100 million views.

That’s less than half a percent.

Production is massive, but success is rare. This explains why AI’s share began to decline in the second quarter of 2026: total output rose to 239,000 productions, while the share of AI-generated works fell to 64%.

A new market: renting out faces

As the industry has expanded, a market that didn’t exist before has emerged in China: licensing human likenesses.

Anyone can offer their face on a specialized platform and specify the types of productions in which their digital likeness may appear. Prices range from $7 to $15,000, according to the South China Morning Post.

Production companies choose the right face based on age, gender, and features, as if browsing an image library.

Retired actors have entered this market, too. Hong Kong actor Lawrence Ng, 62, licensed his likeness as it appeared when he was 20 for use in an AI-generated film.

Even death is no longer the end of the journey. American actor Val Kilmer, who died in 2025, will return to the screen in “As Deep as the Grave” in a fully AI-generated version. He had agreed to take the role before his death but was unable to film any scenes. The filmmakers recreated his image and voice digitally with the consent of his family and heirs, according to Variety.

When a face is used without permission

The market is not without disputes.

Two influencers accused Hongguo, a ByteDance-owned platform, of using their likenesses in a drama without their consent. The platform took the production down and suspended the producer from uploading content for 15 days.

The platform says it has since reviewed 15,000 productions for improper AI use and taken action against 670 of them. ByteDance says it has removed more than 85,000 clips involving unauthorized reproductions of people’s faces and voices since the start of 2026.

Lawyers warn that a single vague clause in a contract could make a face available for any use, anywhere, forever.

Who loses out?

The industry’s workers are paying a steep price.

In China alone, around 690,000 people working in the short-series industry are at risk of losing their jobs, according to the Financial Times. They include not only actors, but also camera operators and lighting, set, and sound technicians.

Production hubs in cities such as Zhengzhou, Xi’an, and Hangzhou—built up during the boom of 2024 and 2025—are now facing a sharp decline in the number of productions being filmed.

But the market is correcting itself

Six major platforms, including Hongguo, Tencent, Mango, and Kuaishou, have pledged at least 6 billion yuan to support short series filmed with real actors. Hongguo alone has earmarked five billion.

Regulators have required AI productions to undergo the same review as productions made by people, and platforms have removed tens of thousands of low-quality works.

One technical problem remains unsolved: keeping a character’s appearance, movement, and background consistent from one scene to the next.

What about us?

The wave has reached us.

The Jeddah Society for Culture and Arts has launched the “AI Films” initiative, enabling creators to use these tools in film productions as part of the goals of Vision 2030.

Saudi experiments have emerged, most notably “Muzna,” a short film by media personality Bader Al Zaidan. It runs for six minutes, stars a character created entirely with AI, and garnered more than 10 million views in just a few days.

Notably, the character has an official account on X, where it interacts with audiences and shares “behind-the-scenes” photos—a virtual-celebrity phenomenon, but with a local twist.

The takeaway

AI has slashed production time and costs like never before, opening the door for people who couldn’t afford to make a single film to produce entire series.

But the questions it leaves behind are harder than the numbers it delivers.

Who owns your face and voice after you sell them? What happens to the hundreds of thousands of people whose livelihoods depend on this industry? And will audiences remain attached to human stories, or settle for what algorithms produce?

Perhaps the half-percent success rate offers an early answer: abundance alone doesn’t make an industry.