Algorithms, addictions and new soap opera age

From LA soundstages to Shanghai studios, microdramas are rewriting how stories are made

French-Taiwanese producer Vincent Wang and line producer Kyle Vertin talk while filming ‘Love Through All Seasons’ at the Wrensmoor Castle in California. Photo: AFP

LOS ANGELES/SHANGHAI:

In a faux castle perched above Los Angeles, cameras are turned sideways, lights are harsh and dialogue is fast. A small crew squeezes into a bedroom to film a raunchy affair scene while a Chinese producer watches closely.

It looks like Hollywood, but not quite. This is the new world of "vertical dramas" - algorithm-driven, smartphone-sized soap operas filmed at breakneck speed and consumed in one-minute bursts.

Once dismissed as kitsch imports from China, these micro-serials have exploded into a multi-billion-dollar industry, merging technology, global production and internet addiction. The format first emerged in China in the 2010s.

"Soap operas on cocaine," Producer Vincent Wang describes the genre that has reshaped viewing habits and is now transforming filmmaking itself - sometimes with no human actors at all.

With budgets of a few hundred thousand dollars and episodes shorter than a TikTok clip, their formula is simple: keep the audience hooked, keep the money flowing. "In 30 days, we can get a show together. Hollywood takes two years. We've already made 500 shows by the time they make their first," Wang said.

The trend began when Chinese amateurs started uploading quick dramas to Douyin, China's version of TikTok. Major companies soon joined in, turning it into an $8 billion business. By 2023, platforms such as ReelShort, DramaBox and FlareFlow had moved into the United States, hiring thousands of out-of-work actors and technicians.

For performers like Zachary Shadrin, who stars in 'Love Through All Seasons', the opportunity feels like survival. "It's the future... right here, right now," he said during filming. After years of Hollywood belt-tightening, the vertical boom offers steady work. But it comes with unsettling undertones.

The genre's most popular titles - 'Dominated by My Dad's Boss' and 'Mated to My Savage Alpha' - rely on melodramatic and often violent plots. "I personally think it's toxic," Shadrin admitted. "But this one was different. It's sweet."

Actor Nicholas McDonald said the shoots are intense but professional. "We all chuckle at some of the lines that are absolutely ridiculous, but everyone treats it very seriously. Because there's money behind it."

Producers know their real competition isn't HBO or Netflix, but TikTok and Instagram. The first dozen episodes are free; then viewers must pay to unlock more. Every sixty-second episode has to end with a twist dramatic enough to stop users from scrolling away.

Director Weiyang Li summed up the formula: "The audience can be hooked right away without using too much of their brain. Everyone's exhausted from their life already."

Production mirrors that urgency. Scripts are translated from Chinese hits, often written on set and sometimes generated by artificial intelligence. Screenwriter Zhiyuan Qu said AI helps suggest ideas and speed up writing. The vertical format also saves money - less background on screen means cheaper sets and smaller crews.

In Asia, these dramas dominate viewing platforms; in the United States and Europe, audiences pay premium rates. Most productions remain non-union, but the Screen Actors Guild recently created a special contract to let members work on microdramas. "I can ditch my side gigs and act again — that's cool," said McDonald. "Eighty per cent of my auditions now are for verticals."

While Hollywood adjusts to this compact storytelling, China is pushing the concept even further - into artificial intelligence. One of the country's latest sensations, 'Strange Mirror of Mountains and Seas', features dragon-like beasts, handsome heroes and entire scenes built by generative software.

With more than 50 million views, it exemplifies the rise of AI microdramas - soap-opera-style shorts sometimes lasting just 30 seconds, created almost entirely by algorithms. Their creator, Chen Kun, said microdramas are the perfect testing ground for AI because audiences watch on small screens and overlook technical flaws.

His production used ChatGPT to write dialogue, Midjourney for still images, China's Kling to animate them, and Suno for the soundtrack. Only editing and voice-acting required people. Fans noticed improvements. "Many special effects can be created using AI," one viewer said on Kuaishou, "and the character expressions are less stiff than before."

AI's impact reaches classrooms too. At the Shanghai Vancouver Film School, teacher Odet Abadia demonstrates tools that can create entire storyboards from a few words. "It's another way of storytelling," she said. "You can get a wow factor, a lot of crazy things, especially in short dramas." Her students experiment with Alibaba's Qwen software, which can instantly produce a plot outline about a wedding photographer entangled in crime.

Abadia insists, however, that students still "shoot with humans and equipment, because we want to support the industry." Yet she admits the future will demand fluency in AI tools.

The technology has already unsettled professionals worldwide. In Hollywood, fears over AI-generated content were central to the 2023 writers' and actors' strikes. The launch of an AI "actress," Tilly Norwood, this year renewed anxieties about performers being replaced.

In Shanghai's live-action sets, those effects are already visible. Artist Louis Liu said AI programs have replaced most concept illustrators. "When AI first emerged, people in film were saying this would spell the end for us. The products were so realistic and cheap."

Still, creators like Chen Kun remain optimistic that new work will emerge in fields such as "prompt engineering," where specialists craft detailed text instructions for generative software. He supports compensating artists whose work trains AI models but says responsibility lies with the original developers, not independent users.

Load Next Story