FestivalComedy

PLASTIC EROSION

by Yifei Li & Zheng Bao

A fish made of plastic travels from a fisherman's catch through market and kitchen to a dinner plate.

Format
Short film
Genre
Environmental satire
Year
2026
Added
2 Nov 2026

Why this film matters

PLASTIC EROSION approaches ocean plastic through one image rather than familiar shots of floating waste: plastic fish that pass along the same chain as food, from the catch to the plate. ITU’s announcement says the filmmakers designed the images to attract at first and disturb on a second look, so the metaphor carries the argument.

The film won ITU’s AI for Good Film Festival 2026, chosen from more than 1,300 submissions and announced at the AI for Good Global Summit in Geneva.

Production evidence

No official standalone full release was found. The Korean awards upload cannot be embedded and its relationship to the ITU version is unclear, so it is listed as a source only. The film is listed as a verified festival screening with official credits. The journal has not watched the complete film.

Production notebook

What the makers have disclosed. Credits may be partial.

01

Credits and result

ITU credits Yifei Li and Zheng Bao (China) as directors and names the film the AI for Good Film Festival 2026 winner, announced on July 9, 2026 in Geneva.

aiforgood.itu.int source ↗
02

Makers' intent

ITU quotes the filmmakers' submission: they started from the question of where discarded plastic eventually goes, and designed images to look appealing at first and disturbing on a second look.

aiforgood.itu.int source ↗
03

Earlier award

The Gyeonggi Content Agency's channel presents a 1:06 work titled Plastic Erosion, credited to Li Yifei (2025, China), as first prize in Art & Culture at the 2025 Korea AI Content Awards. Its synopsis says the whole production, from concept and script to modelling, rendering and set design, was done with AI. Whether that is the same cut ITU awarded is not established; ITU lists 2026 as the production year.

YouTube source ↗
04

Unknown production details

Specific models, tools, runtime of the ITU version and budget are not disclosed.

aiforgood.itu.int source ↗

Open at source ↗