REAL PLASTIC: An Experimental Study

REAL PLASTIC is a series of fictional documentaries exploring the poetic and surreal afterlives of mannequins. This project seeks to capture the beautiful and haunting echoes of an era left behind, using the very technology that is currently redefining our future.

I began this study by looking at the mannequin as one of our first "fake" human. Decades ago, we created these static, plastic ideals to replace real bodies in our shops and our lives. They were inanimate objects, yet they felt strangely alive within the cultural memory of the American mall. Today, as digital spaces and AI-driven avatars render physical retail—and many human roles—obsolete, these mannequins have become out-of-work relics.

I felt that AI was the perfect medium to bring these "dead" objects to life. There is a deep irony in using a new "fake" technology to resurrect an old one. It allowed me to explore our current fears about AI replacing human workers by looking back at the mannequin, which once did the exact same thing in the analog world. My goal was to see if I could use these tools to find a soul in the synthetic and a sense of movement in what was once frozen.

This project was also a study in production efficiency and solo creative direction. I set out to create a complete, high-fidelity cinematic experience in just five days. Every visual, every movement, and every sound was directed and executed by me, demonstrating how AI can turn a single artist into a fully functioning film crew.

The Workflow

To produce this film within a 120-hour sprint, I developed a rapid generative pipeline that moved from static composition to cinematic motion:

• Composition & Mockups: The process began in Adobe Photoshop, where I created mockup files to establish the lighting, framing, and the specific atmosphere of the abandoned mall and the city bus.

• Neural Generation: For the character and environment, I utilized a combination of ComfyUI (Flux model) and Nano Banana (Gemini 3 Flash Image). To ensure the mannequin felt consistent across different scenes, I integrated several custom LoRAs to capture the exact material texture of 80s plastic and the era's versatile aerobic and classic fashion.

• Cinematic Animation: To give the mannequin her eerie, model-like walk, I used Gemini Veo 2 within Flow, the browser-based animation tool. This allowed me to transform my static frames into fluid, cinematic video sequences with high temporal consistency.

• Resolution Scaling: The final output was processed through a custom Flux upscaler in ComfyUI, bringing the fine details and textures into a professional, high-definition resolution.

• Aural Design: The narrator is my own voice. I manipulated the recording using Adobe's professional audio tools (Audition), altering the tone and frequency to create a pensive, observant voice—someone who is kind and non-judgmental, but sees the world with a precise, historical point of view.