How Two Non-Filmmakers Won $100K with an AI-Generated Short Film

The barrier to creating compelling video content has just been shattered. Forget film school, expensive equipment, and large crews. The latest proof comes from two unlikely creators who just won a top prize—and $100,000—in a major AI video competition, all without a traditional film background.

One is an advertising professional with a 3D animation degree, and the other is a biology PhD student born in 2001. Their winning short film, titled “Brand,” was created in just 23 days. It racked up over 10 million views, 800,000 likes, and 300,000 shares on Bilibili within a week, even earning praise from Guo Fan, the acclaimed director of “The Wandering Earth.”

This story isn’t just about a viral video; it’s a blueprint for the future of creative work. Let’s break down how they did it and what it means for anyone looking to harness AI for storytelling.

The Winning Formula: Clarity Over Clever Prompts

When you see an impressive AI-generated film, the first instinct is to ask: “Which model did they use?” or “Can I have the prompt?” The assumption is that a magical tool or a secret command is the key. However, the winners of Bilibili’s first AI Creation Contest, which saw over 8,300 entries, succeeded for the opposite reason.

Their process was remarkably disciplined and human-centric. They didn’t start with the AI; they started with a foundational creative question.

Step 1: The Brutal Question of “Why?”

DiDi_OK, winner of the “Open Track,” begins every project with a harsh self-interrogation: “Is this idea necessary to express?”

He argues that an idea must be so urgent it keeps you up at night. Only then does it pass the first filter. The second filter is audience-centric: “Will this expression bring value or just be a burden to others?” The goal is to offer something new and interesting.

This focus on core expression is non-negotiable. He quotes filmmaker Jiang Wen: “I made the dumplings just for that vinegar.” If the central idea (the vinegar) gets lost in the production process (making the dumplings), the entire project isn’t worth pursuing.

Step 2: Structure is Your AI’s Best Friend

With a solid “why” established, the next step is designing a simple, understandable form. DiDi_OK likens making a film to hosting a dinner. Your core idea is a strong, personal dish. You can’t just shove it in the guest’s face. You need to set the table, create the mood, and guide them to the main course at the right moment.

This is where traditional filmmaking structure becomes your greatest asset against AI’s randomness. Bill_, winner of the “Three-Body Problem Track,” embodies this engineering mindset. He refuses to be led by the tool. Instead, he inserts AI into a classic animation pipeline:

1. Script & Storyboarding: He hand-draws rough “stick-figure” storyboards on an iPad to lock down composition, character placement, and action.
2. AI Generation as a Render Farm: He feeds these structured sketches to image models. This provides a much stronger constraint than vague text prompts, guiding the AI toward a specific vision.
3. Layered Prompting for Video: For video generation, his prompts are meticulously structured in three layers: the scene’s overall tone, precise details from the script, and finally, atmospheric and character nuances.

The Art of De-AI-fying Your Content

A common pitfall of AI video is the uncanny, synthetic “AI look” that pulls viewers out of the story. The winners shared their tactics for overcoming this.

Be Specific, Not Vague: Don’t prompt for “anime style.” Instead, reference the specific visual language of a known work (e.g., “the color palette and line work of Studio Ghibli’s Spirited Away“).
Use the Language of Film: Direct the AI using cinematic terms. Control the camera angle (low-angle shot for power, high-angle for vulnerability), plan a color grade journey, and design shot sequences.
Design Expressive Characters: Use characters with exaggerated, lively expressions and clear emotions. This tricks the viewer’s brain into believing there’s a human performer behind the pixels, fostering empathy.
Sound is 50% of the Experience: Many AI shorts have stunning visuals but empty, flat audio. Bill_ uses sound design—like a rising低频 hum or a carefully placed sound effect—to build tension. DiDi_OK, who admits to being “tone-deaf,” used AI music tools like Suno by first having an LLM analyze the cultural roots of a reference track (like the folk vocals in The Witcher 3 score) and then turning that analysis into a rich, thematic prompt.

Tool Mastery: Knowing Your AI’s Personality

The creators don’t rely on a single “best” model. They treat each AI as a specialist with a unique personality and assign tasks accordingly.

For Concept Art: Use Midjourney for initial artistic vision, then refine details with a tool like Nano Banana.
For Video: Google’s Veo is praised for superior lip-sync control. Kling excels at creating a gritty, documentary-like texture with heavy,细腻的镜头感. Seedance 2.0 is noted for its authentic handheld camera shake.

The lesson is clear: Master the workflow, not just the widget. The flashy AI generation in the middle is just the execution layer. It’s the solid creative foundation before it and the rigorous editorial judgment after it that create great work.

The New Creative Reality: AI Accelerates the Journey

This shift has profound implications for all creators.

1. “Being Yourself” Becomes a Strategic Advantage

AI dramatically lowers the cost of experimentation. As DiDi_OK puts it, you can now quickly produce your “first ten garbage works”—a concept from the game design bible The Art of Game Design. The idea is to rapidly iterate through early, imperfect versions, get feedback, and improve. Your unique voice and perspective are no longer a liability but the primary differentiator in a sea of AI-generated content.

2. The Focus Shifts Back to the Human

When a powerful new AI model drops, there’s a brief frenzy of creating visual spectacles—giant monsters, epic disasters. But as semi-pro Bill_ notes, once everyone can do that, the novelty wears off. “When a hammer becomes so handy you don’t have to think about it, you stop studying the hammer and start looking at how well the nail is driven,” he says. The audience’s attention returns to story, character, and meaning.

Director Yi Xiaogang states it bluntly: “AI can’t write comedy because it doesn’t make mistakes. And it’s precisely ‘mistakes’ that make artistic works captivating.” Your taste, your cultural knowledge, and your understanding of humanity are what ultimately decide a project’s success.

3. Pre-Production is More Critical Than Ever

AI doesn’t replace the need for planning; it demands better planning. Writers can now describe scenes they previously thought were unfilmable. However, this requires a deep understanding of the model’s strengths (spectacle, action) and weaknesses (subtle facial acting, long dialogues). You must design your script and workflow with these boundaries in mind.

DiDi_OK views AI as a black box. You can’t control its internal workings, so you must be exceptionally clear with your input and ruthless with your output selection. Creativity, scriptwriting, and aesthetic judgment have moved to the very front of the line.

Your Move: Start Creating Now

The $100,000 prize might seem distant, but the tools and opportunities are here today. Bilibili has announced long-term support for AI creators, including funding and traffic incentives through programs like its “AI Animation Theater.” They’ve even launched updream, an AI assistant designed to learn and adapt to an individual creator’s style and preferences.

The message from these pioneers is empowering. The excuses of budget, team, or technical skill are fading. The most important step is to start. Go make your first “garbage” project. Then make another. By the third one, you might just find your unique voice—amplified by AI—is exactly what the world is waiting to see.