Original · Unique Research · 2026-07-08
Editor's note: This is the original Chinese author's interview with Luo Yuan and its framing. This English rendition retains the full text in source order, including all 12 Q&A items. The interviewee's views, financial figures, and market claims are her self-reports, attributed to her and not independently verified. Person, company, and work names are preserved as source attributions.
AI Industry Observation
When everyone uses AI to speed up production, what is the thing that will truly remain?
The AI short-drama industry lacks nothing but speed right now. One person, two days, one short drama, overseas views breaking ten million — over the past year such stories have been endlessly cited as proof that AI is rewriting the content industry.
But the real question is: when everyone can generate images faster, produce video more cheaply, and test content more frequently, what is left behind?
Luo Yuan (罗媛), founder and executive director of Phantasm Era (幻境纪元) — the "Shanzhu" (山竹) many people know — gives a direct answer: emotion. The laws of content. And, don't treat the audience as a fool.
"Tools can be standardized; emotion can't."
She doesn't quite believe that "AI tools automatically bring hits." On the contrary, she thinks AI makes passing-grade content easier, but also raises the bar for content truly worth remembering.
From Super Creator to Founder, She Hasn't Left the Creative Floor
Luo Yuan now has two public identities: founder and executive director of Phantasm Era, and an AIGC multi-platform super creator. The two sound like one company-side and one content-side. But in her case they aren't split. She now spends most of her time on company management, but keeps a high proportion of creative involvement. Her time split: 60% creation, 40% management.
This ratio is telling. Many creators, once they start a company, quickly retreat from the content floor to the management back office. But Luo Yuan hasn't put down the "super creator" label. The reason is practical: "Most of the company's partnerships come from my craft and creation." Her creative ability isn't decoration on a business card; it's a key source from which Phantasm Era wins partnerships and builds commercial trust.
In the past she could pull off "one person, two days, one overseas short drama online." Now this extreme personal way of creating is becoming a team effort. She says she has a creative team now, and her own role is more like "producer." Behind this is a migration of an AI-content company from personal hit-making ability to organizational production ability. One person can run a hit, showing the method works; but to keep doing content, business, and IP, the method must enter the team.
After the team joined, her creative style also changed. Not more industrialized, not more assembly-line, but "more refined." More importantly, broad team opinions and discussions produce new ideas. This is a realistic dividing line in AI content startups: early on, personal talent and tool dividends can carry you through, but to go further you can't rely on one person's feel.
She Doesn't Want Safety; She Wants the Audience to Remember
Luo Yuan dropped a hard line:
"I'd rather make you feel offended, angry, or sad, as long as you remember my short drama."
In the content industry, this line is actually accurate. The problem with much failed content isn't that it's bad, but that it's too flat. No offense, no conflict, no pain, no memory point. The audience scrolls past and nothing stays. Luo Yuan's own hit shorts are basically all tied to strong emotion. She mentions "Jiecao Xianhuan" (结草衔环), "Gu Shi" (蛊誓), and the recent "Langya Mountain" (狼牙山), all built on stirring audience emotion. Works she herself feels "nothing" about usually perform poorly.
This judgment isn't new to the AI era. Film, short drama, novels, comics are essentially the same. The audience won't lower their story requirements just because a video is AI-generated. On the contrary, as AI content grows, the audience faster recognizes what's an empty shell versus what truly has emotion, expression, and attitude.
"Gu Shi" is a typical case. This short doesn't stop at "two people simply falling in love"; it adds a war background, patriotic feeling, and life-and-death character choices. Luo Yuan says this is what she naturally wanted to express after picking up creation tools as a creator: "I wanted to express something different, to let the audience sense, even awaken." AI's role isn't to decide what she expresses, but to speed up execution. She first uses AI to make the content, then supplements, adjusts, and refines. Without AI tools, the cost and cycle of such a film would be completely different. She says just location scouting and shooting "would take a good month," and several locations might have to change. This is AI's most realistic value to content creation: it lowers the bar of expression but doesn't replace "why express it."
Tools Are Standardized; Emotion Isn't
Now many teams in the industry use similar AI video tools. Kling (可灵), Jimeng (即梦), Runway keep iterating; image quality, consistency, and duration all improve. The question follows: when everyone's tools look more alike, will content look more alike? Luo Yuan's answer is simple: "Emotion is made of the public's and the individual's experiences and encounters. Everyone's emotional experiences and encounters differ; even if the same, the way they're expressed differs." Her emotional moat, in the end, comes from "life experience."
This answer isn't complex, but it's key. AI can learn film language, simulate art style, generate characters and scenes, and help people complete narrative structure faster. But it's hard for AI to decide for a creator: why were you actually moved by something? Why tell this story? Why should a character live here and die there? Why is this moment worth magnifying? These things can't be simply prompt-engineered. So Luo Yuan always stresses that what AI produces is standardized, while emotion is non-standard. Standardized tools let more people enter content, but what truly decides the difference is still the creator's understanding of people, relationships, fate, and emotion.
Her operations across platforms follow the same logic. She now creates mainly on Douyin, Xiaohongshu, and similar platforms. The core is still emotional density, but expression adjusts to platform tone. Douyin is more entertainment-oriented, suited to cheerier, more mass-market video; Xiaohongshu is more emotional and better suited to practical content and women-related content. She observes that with the same content distributed out, Douyin data usually beats Xiaohongshu. Douyin prefers cheerful content; Xiaohongshu eats more for a women's perspective and emotional expression. This shows the so-called content method isn't one template for all. The underlying law can be shared, but platform context must be adjusted.
AI Hasn't Changed the Basic Laws of the Content Industry
At the April Hangzhou Unique Awards roundtable, Luo Yuan offered a judgment: "The content industry's laws won't change because of AI. Emotional density, story structure, character dimensionality — these hold in film, hold in live-action short drama, and still hold in AI comic drama." Three months on, she hasn't revised it. She still believes these must exist. This is also her core logic on the loss problem in the AI short-drama industry.
As for the claim that "90% of AI short-drama companies lose money," she agrees. Why so many lose money? Her judgment isn't "poor tool use" but "not understanding content laws." This line is important. Over the past year, there's been a strong technical hallucination in AI-content startups: as long as you can make the image, as long as capacity is high enough, as long as ad placement and distribution run, the content holds. But reality is that the audience won't pay just because you used AI. Luo Yuan says Phantasm Era stays in that 10% by doing one thing right: "Make valuable work and content; don't treat the audience as a fool."
"AI can raise efficiency, but it can't replace respect for the audience."
Many AI short dramas' problem isn't crudeness, but contempt for the audience. Treating viewers as a traffic pool only stimulated by beauties, reversals, spectacle, and dopamine hits; making content into skin-change templates; making emotion into cheap buttons. Short-term data may come, but long-term brand, IP, and repeat purchase are hard.
She Chooses Refinement + IP Over Volume Gambling
The AI short-drama industry now roughly has two routes. One is refinement and IP-building: signing top authors, strong settings, strong characters, strong worldview. The other is volume betting on hits: batch production, fast testing, gambling on probability. Luo Yuan places herself in the former: refinement + IP. Her differentiation isn't one ability but several stacked: creation speed, emotional intensity, and platform operation.
This also defines Phantasm Era's business model. Stable revenue sources she mentions include platform revenue sharing, self-media, and commissioned short-drama production. Also, from April to now, AIMV-related partnerships have grown. She observes MV is also moving toward AI. This shows commercial demand for AI imaging content isn't only from short dramas themselves; it's spreading to music visuals, brand content, and custom imaging. On overseas short dramas, she doesn't simplify it to "go overseas to grab traffic." She mentions the team is moving overseas and has strong overseas hit-making experience, while having also figured out Douyin self-media creation logic. To partners, what they're valued for includes three things: hit rate, visual effect, and content tone. These three words together are basically the core commercial metrics of an AI-content company. Visual effect alone isn't enough; platform running alone isn't enough; content tone alone without hit rate is hard to turn into a steady business. What truly holds is the combination of creative judgment, technical execution, and platform operation. As for model costs and video-generation compute rising, her answer is restrained: there's currently compute support. This also reminds of a real issue: an AI-content company's cost structure is still tightly tied to compute, tools, and model platforms. Short-term compute support does help the team run faster; long-term, whether it settles into content assets and commercial barriers still depends on the work itself.
Low-Quality Volume Will Disappear; AI Skin-Change Will Die
On the future, Luo Yuan's judgment is clear. If she could only go all-in on one direction, she'd still choose refinement. And she thinks the future may gradually converge on film and television drama. This means she doesn't see AI short drama only as a short, flat, fast traffic business, but as a content form that may enter longer narrative, higher production specs, and stronger IP. A year later, which of today's hot plays will disappear? Her answer: batch production of low-quality short dramas. Which seemingly cold plays will be proven right? Her answer: persisting in refinement and content creation. As for content purely skinned with AI, her judgment is blunter: it will die. What may survive is live-action + AI hybrid, strong-IP-driven, and celebrity-effect content forms.
"Technical dividends will disappear, but content assets remain."
Early on, people are drawn to "AI made it." But as AI video becomes common, the audience's excitement shifts fast. They won't pay long-term for "this is AI-generated"; they'll pay for "this story is worth watching."
The Knife Got Sharper; the Craft Didn't Depreciate
Luo Yuan once said: "Technology gave everyone a knife, but the knife itself can't cut meat." Now the knife is indeed getting sharper. AI video generation quality, consistency, and duration all improve. Making passing-grade content gets easier. Has the craft of "cutting meat" depreciated? She thinks making 90-point content got easier. At the same time, passing-grade content flooding in also raises the bar for truly good content. This isn't contradictory. What AI lowers is execution difficulty, letting people with expression more easily make what's in their head; but it also raises market noise, making ordinary content easier to drown. So two outcomes will appear: on one side, lots of homogeneous, low-quality, skin-change content piles up fast, then is eliminated by platforms and audiences. On the other side, people with real content judgment, emotional expression, and aesthetic ability use AI to make more valuable work. AI hasn't made content creation simpler; it's made content stratification faster.
If She Could Start Over, She'd Build a Team Earlier
From super creator to founder, Luo Yuan looks back and would do one decision differently: build a team earlier. This line is representative. Many AI creators early on go through a stage where one person can finish what used to need a small team, producing a strong personal-production hallucination. But when content moves from single hits to continuous operation, from personal accounts to company-scale running, from works to commercial delivery, one person's ceiling shows fast. The team isn't to replace the creator but to amplify them. Especially in AI content, a team's value isn't just division of labor but discussion, collision, review, and iteration. One person's emotion and experience is a moat, but a company can't rely only on one person's in-the-moment state. So Luo Yuan is now more a producer and a content operator. What she does isn't finish every step by hand, but turn her content judgment, aesthetic standards, and platform experience into a capability the team can keep executing.
Making Content in the AI Era: Start With the Small Things Around You
If she could give AI entrepreneurs, content creators, and investors only one piece of advice, Luo Yuan says: "Make valuable content; the essence of making money is how much help you give people and society." This differs from many people's imagination of AI content. The AI-content industry too easily falls into technical narrative: models, tools, prompts, workflows, costs, efficiency, capacity. But Luo Yuan finally pulls the question back to content itself: what exactly do you give others? What do you let them see? What do you let them be touched by? Do you help the audience understand themselves, understand relationships, understand some situation? Her concrete method is simple: start observing the small things around you — what are the small things that truly move people and resonate. This may be her biggest difference from many AI short-drama practitioners. She doesn't treat AI as the answer to content, but as a tool of expression. The tool handles acceleration; the creator handles judgment. The tool handles generation; the creator handles choice. The tool makes the image, but whether to make someone angry, sad, offended, or struck is still the creator's job.
The real competition in AI short drama may never be who uses tools better, but who understands people better.
Selected Interview Q&A
Q1: You're now both founder and executive director of Phantasm Era and an AIGC multi-platform super creator. How do these two identities split your time?
Luo Yuan (Shanzhu): Now it's mostly company management, but I also carry the super creator's technical ability. Overall, creation is about 60%, management 40%. Actually a lot of the company's partnerships come from my technical and creative ability. So the "super creator" label still holds for me, and now it covers more ground.
Q2: From "one person, two days, one overseas short drama online" to creating with a team now, how has your role changed?
Luo Yuan (Shanzhu): There's a creative team now; my own role is more like producer. With the team, creation definitely changes. Overall it's more refined, and broad opinions and discussions also produce new ideas. Before it was more one person's extreme creation; now it's the team together polishing content more completely.
Q3: You once said, "I'd rather make the audience offended, angry, or sad, as long as they remember my short drama." How did this creative creed form?
Luo Yuan (Shanzhu): My hit shorts basically all stir emotion more or less. Like "Jiecao Xianhuan" and "Gu Shi," each with ten-million views, and the recent "Langya Mountain," all built on stirring audience emotion. Some works I myself feel nothing about get poor traffic. So I think content must give the audience emotion — maybe anger, sadness, or being touched — but it can't be flat.
Q4: "Gu Shi" added war background, patriotic feeling, and life-and-death choices; why not a simpler love story?
Luo Yuan (Shanzhu): This is what I as a creator naturally wanted to express after picking up creation tools. I wanted to make something different, to let the audience sense, even have a little awakening. AI tools mainly speed up my execution — first make the content, then I supplement and adjust. Without AI, this film's cycle would be much longer. Location scouting and shooting might take a month, and several locations might change.
Q5: You said "what AI produces is standardized; emotion is non-standard." When everyone uses similar tools, how do you avoid content homogenization?
Luo Yuan (Shanzhu): I think emotion is made of the public's and the individual's experiences and encounters. Everyone's emotional experiences and encounters differ; even with the same experience, the expression differs. If my "emotional moat" is anything, I think it's more life experience. AI can help you generate things, but it can't live your life for you, nor decide why you're actually moved by something.
Q6: You now distribute content on Douyin, Xiaohongshu, and other platforms. How do content strategies differ?
Luo Yuan (Shanzhu): The underlying logic is the same — still following emotional density — but concrete expression differs a bit. Douyin is more entertainment-oriented; as people say lately, it's more about entertaining the masses. Xiaohongshu is more emotional and better for practical, women-related content. We distribute normally, but you clearly feel Douyin likes cheerful video, Xiaohongshu prefers women-related content. Overall, Douyin data beats Xiaohongshu.
Q7: You mentioned before that the content industry's laws won't change because of AI. Do you still hold this view?
Luo Yuan (Shanzhu): No change; I still feel they're needed. Emotional density, story structure, character dimensionality — these hold in film, in live-action short drama, and in AI comic drama. AI tools can raise efficiency, but content's basic laws won't disappear just because AI appeared.
Q8: Do you agree that "90% of AI short-drama companies lose money"? If so, where do you think they lose?
Luo Yuan (Shanzhu): I agree. I think many companies don't lose on poor AI tool use, but on not understanding content laws. We keep going because we insist on making valuable work and content, not treating the audience as a fool. The audience can actually feel whether you're seriously making content, whether you really want to express something.
Q9: What's Phantasm Era's main business model now? Where does steadiest revenue come from?
Luo Yuan (Shanzhu): Mainly platform revenue sharing, self-media, and commissioned short-drama production. From April to now, there are also more AIMV-related partnerships. Now MVs are moving toward AI too, so this demand is growing. We currently run the refinement + IP route, not pure volume betting. In differentiation, we have creation speed, emotional intensity, and platform operation.
Q10: Looking back at the AI short-drama/comic industry a year from now, which plays do you think will disappear? Which directions will be proven right?
Luo Yuan (Shanzhu): I think batch production of low-quality short dramas will disappear. What will be proven right is still persisting in refinement and content creation. Pure AI skin-change content will die. What survives may be live-action + AI hybrid, strong-IP-driven, and celebrity effect. If I could only go all-in on one direction, I'd still choose refinement, and later it may gradually converge on film and television drama.
Q11: AI tools get stronger; passing-grade content gets easier. Is making 90-point content harder or easier?
Luo Yuan (Shanzhu): Easier. Technical progress lets creators more easily make ideas. But after passing-grade content floods in, it also raises the bar for truly good content. So there will be more content, but truly valuable content will also be easier to see.
Q12: If you gave AI entrepreneurs, content creators, and investors one piece of advice, what would it be?
Luo Yuan (Shanzhu): Make valuable content. The essence of making money is how much help you give people and society. More concretely, start observing the small things around you — what small things truly move people and resonate. Content must return to people, to real emotion and resonance.