This is the fifth episode of my podcast. Jasmine graduated from an Ivy League university in the US, spent the past six or seven years helping Chinese consumer brands go to Southeast Asia, and has been deeply involved in the content industry β investing in one of China's top-tier IPs, and this year jumping into the frontlines of AI + film production.
This episode wasn't about technology. It was about people. Jasmine discovered that the tech world and the creative world hold two extreme positions on AI β but privately, both sides are contradicting themselves. She sits in the middle as mediator, pressing people who can't stand each other into the same room. She says the hardest problem in the AI era isn't technical. It's human.
The two 90-percents
While sourcing AI content generation teams, Jasmine asked every founder the same question: can AI replace content creators?
About 90% of tech founders told her: directors and screenwriters will no longer be needed. Everything can be done by technology. They were supremely confident.
Ask the creators, and 90% β even 95-99% β said: AI is nothing. Self-expression, storytelling β only humans can do that. Technology can't replicate it.
But privately, neither side lives by what they preach. When Jasmine reviewed work from "fully AI-generated, no human intervention" tech teams, she'd ask: "Is this really made with zero human involvement?" They'd admit they still heavily relied on people to calibrate and control the storytelling. And those directors and screenwriters who publicly dismissed AI? Privately using it to write scripts and generate pre-production concepts.
Both sides are lying to themselves. Her job is to get everyone past the posturing and into the same room to actually work together.
"80s feelings" β Her job is Chinese-to-Chinese translation
Tech teams think creative people are vague and impractical. Directors say things like "the vibe isn't right" or "this isn't the feelings I want," and engineers are baffled.
A real example: a director said he wanted "80s feelings." The engineer's response: "So you want me to add a retro filter?"
The director's creative pride kicked in β the engineer just doesn't get it. But what Jasmine can do is visualize β take the director's abstract sense and make it concrete: 80s feelings from mainland China, Hong Kong, or America are completely different visual languages. Which one? Once the tone is set, lighting is one piece, the characters' spiritual state is another, facial expressions another.
What she does is "Chinese-to-Chinese translation" β taking a creator's abstract expression and converting it through visualization and quantification into something a tech team can actually build toward.
Generation vs. creation β a subtle but critical distinction
Jasmine has been doing writing exercises, comparing her own work against AI output. Her conclusion: for now, AI is "generating," not "creating."
Give AI a prompt with full context and it can expand, it can continue a narrative based on what you've fed it. But it follows existing material and training data. It generates. It doesn't create.
She verified this by forcing herself to watch AI-generated comic dramas that are performing well in clicks. She could keep watching, but what engaged her was plot progression β like listening to audiobooks as a kid. She wasn't moved by the visual storytelling itself.
Traditional directors use wide-medium-close shot transitions, lighting, camera angles β you can empathize with characters beyond just the plot. Current AI comic dramas can't achieve that interactive process.
"Using the word 'replacement' at this stage is genuinely premature."
Intervention, influence, replacement β three very different words
Jasmine draws precise lines between these three concepts.
Replacement β she disagrees. AI can't tell a story well yet.
Influence β her ideal state. AI does information gathering during project initiation, concept design during pre-production, process optimization during shooting. It exists as a facilitator. The final decision-maker is still the creator.
Intervention β the negative example. She described a manga adaptation project where Korean comics needed localization with redesigned world-building and values. The team used AI to assist, but the AI's training data was Western-centric, creating major conflicts with the creator's vision.
The young team, strong on internet instincts, pushed back: "AI represents the majority feedback from users, and it disagrees with my idea. What do I do?"
Jasmine's judgment was clear β internal consistency matters. If your premise, development, and conclusion have logic, that's enough. But the team got stuck between their own instincts and AI's majority-vote feedback. A project that shouldn't have taken long got dragged out because AI's participation extended the pre-production phase.
The hardest problem in AI isn't technical β it's human
Managing the fragile ego of artists is part of Jasmine's job. In this wave of AI disruption, the problem isn't the technology.
Creators' resistance to AI isn't a rejection of technology β it's fear of being replaced. When you discuss problems from a place of fear, you're not focused on the problem itself. You're self-protecting. People reflexively argue back, reflexively dismiss.
Her method is incremental. When AI teams and content teams sit together, she strictly prevents the conversation from sliding toward "can AI take over human roles." Focus on small tasks β start with a specific product concept, let AI provide some directions, let the team discuss on top of that.
A real case: an IP's product design team initially insisted on making extremely hardcore, precision-crafted, high-ticket collectibles. But through iterative exchange with AI, they realized this would only satisfy a tiny circle of die-hard fans, while ignoring the emotional assets the IP had accumulated over years β how could ordinary users participate in the IP's story in an immersive way?
AI didn't propose this direction. But AI's imprecise suggestions helped the team identify what they should actually be doing. Most critically β everyone saw that AI genuinely can't do everything, and the creators' anxiety dissolved.
China's IP industry skipped the hand-crafted phase
Jasmine says China's IP development and operations are still at a very premature stage.
"Learn from Disney, learn from Pixar, learn from Marvel" β people have been saying this for years, but China hasn't produced a truly long-lived IP with real vitality. The difference with the US and Japan is that those legacy studios all went through a raw, hand-crafted stage. Japan is still hand-crafting. China basically skipped that phase and jumped straight into AI-assisted production.
AI's biggest help to her personally: she'd always wanted to do IP operations but felt she had more important business to handle. Now AI can visualize the blurry images in her head, reducing initial costs and lowering the decision threshold. She no longer needs to mobilize an entire design team to sketch her ideas.
Still investing in people β but with one new dimension
As an investor, some of Jasmine's criteria are changing. Others aren't.
What's changed: she no longer wants to invest in projects with 3+ year production cycles and 5+ year payback periods. When she was younger and the economy was booming, you could throw big money at big IPs without worrying. Now the entire industry is quietly shrinking project sizes and shortening payback windows.
What hasn't changed: she still invests in people. "The value of content is permanent. What changes is the medium and the vehicle." She's been through print media, the self-media explosion, film and TV, short dramas and comic dramas, and now AI-generated content. Every phase produced winners. Not because they chased formats β because they believed in people.
But now she evaluates creators with one additional dimension: beyond storytelling ability and attitude toward their work, she looks at their attitude toward technology, their learning capacity, and their ability to collaborate across different teams.
"I know a bit of everything β that's my core competency"
Jasmine is a pure liberal arts person. Double major β International Relations and Economics. But everything she's done β going overseas, consumer brands, content investment, IP operations β is the work closest to people.
In her twenties, she felt insecure about not having a "solid skill." Now her answer is different.
"I know a bit of everything. Being a connector and mediator β that's my core competency."
Her advice for liberal arts students: don't trap yourself in the arts-vs-science label. That division is man-made. Cross-disciplinary ability is the strongest asset in this era. Ground yourself in your own field, then branch out β see how your expertise can connect to other industries. That connection itself is the biggest opportunity for liberal arts people.
*This article draws from EP05 of my Chinese-language podcast on AI founders going global.