Three people, all born after 2000, created an AI-generated mystery-thriller series that has drawn more than 1 million likes per episode on Douyin. The Inverse Zone, an original monster story dubbed an AI version of Stranger Things, shows how small teams are now producing cinematic content that once required full production crews.
The series was created using domestic video-generation models Seedance 2.0 and 2.5. Team member Hu Hai told the Global Times the technology made what was previously impossible suddenly feasible. "We've always wanted to make professional film and television content, with visual language, complete storytelling and a cinematic feel," he said. "But the cost and barriers of live-action production were too high. Small teams had little chance to realize high-quality film and television projects."
Production is not effortless
A typical episode costs around 4,000 to 5,000 yuan ($593-$741) to produce, Hu said. Episodes with large-scale scenes, complex special effects or demanding lighting and rendering runs can exceed 10,000 yuan. The team uses AI for image generation, basic voice work and material rendering, while humans handle the script, character development, cinematography and post-production.
"The soul of the entire series is still controlled by people," Hu said. "All the key creative decisions are manually adjusted and repeatedly reviewed."
Ya Cen, operations head of AI video platform LibTV, described the current ecosystem as small teams with high productivity and strong individual styles. "The most direct value of AIGC is that it has lowered the geographic and financial barriers to film and television production," she said.
Hu's team studied directing and filmmaking in college and initially worked on short comedy videos. After years of producing that content, they hit a creative plateau before more advanced AI tools changed their options. The two-month preparation process for The Inverse Orbit included world-building, character design and establishing a consistent visual style.
Actors feel the pressure
The shift extends beyond production teams. AI-generated characters are building fanbases, landing brand endorsements and being described as the next generation of digital stars. Several celebrities, including Taiwanese actress Joey Wong, have reportedly given permission for their younger likenesses to appear in AIGC.
Luo Xinyuan, a 21-year-old from Chongqing, spent more than a year in Hengdian as a short-drama actress before moving to a scenic-area role playing Huang Rong, a character from classic martial arts novel The Legend of the Condor Heroes. "When short dramas were at their peak, there were lots of fast-paced, feel-good stories," she told the Global Times. "But that kind of content is difficult to sustain. Without richer and more meaningful stories, the pressure from AI short dramas will make things even harder."
Zhang Peng, a cultural researcher and associate professor at Nanjing Normal University, called the disruption to actors' work "structural rather than an immediate wholesale replacement." He warned that unrestricted celebrity likeness authorization could raise control and revenue-sharing concerns, putting further pressure on ordinary actors without clearer rules.
Skills that transfer
For creative professionals watching these shifts from a distance, the technical details of AI video work are becoming relevant to more than just film production. Storytelling structure, character consistency and visual coherence are the same challenges any creative team faces when adopting AI tools. A AI Learning Path for Video Editors covers the generation and editing workflows now central to productions like The Inverse Orbit. Scriptwriters face a parallel adjustment as AI-generated characters require the same narrative discipline as live-action ones - an AI Learning Path for Scriptwriters addresses those workflow changes directly.
Why this matters for creatives
The Inverse Orbit's three-person crew demonstrates that AI closes production gaps that used to require entire departments. But the actor stories show what happens when entry barriers fall from below. The next stage is likely to be neither human-only nor machine-only production, but a hybrid model. For professionals in television, film and online content, the takeaway is concrete: script quality and visual control separate AI-assisted hits from the noise.
Your membership also unlocks:
