About the Project
AIRWAVES: Ripple Effects is an AI-generated short film that examines the relationship between truth, memory, and systemic distortion. Rooted in themes of censorship, extraction, and collective responsibility, the film investigates how centralized systems in politics, media, and technology reshape narratives in ways that obscure the very histories they claim to protect.
The film is built from a multi-year personal archive that includes images, audio, and video collected across regions and platforms. By filtering this dataset through AI tools, the project exposes how embedded bias influences what these systems choose to reveal or conceal. The distortions become a lens through which deeper truths emerge.
Inspiration
The project was inspired by years of observing how media infrastructures filter access to information. Many of the images and sounds in the film originated as documentation of censorship, protest, migration, and daily life across the Global South. Working with closed-source models revealed patterns of erasure and reinterpretation that mirrored the very systems the project critiques.
The piece also draws from ongoing research on digital memory, participatory archives, and algorithmic bias, as well as a long-standing interest in glitch artifacts as carriers of truth.
How the Project Was Built
The film was constructed through a layered pipeline using closed-source AI systems. Each tool contributed its own form of distortion, which became part of the narrative.
Each output was treated not as a final image, but as a data point revealing how the system itself interprets the input material.
Challenges We Faced
- Working with opaque systems that restrict control over outputs
- Managing unpredictable distortions introduced by closed-source models
- Preserving emotional truth while navigating erasure and reinterpretation
- Integrating large amounts of archival material into a short runtime
- Designing sound that communicates suppression, memory, and mechanical rhythm without overwhelming the visual narratives.
Built With
- adobe
- luma
- midjourney
- runway
- stableaudio

Log in or sign up for Devpost to join the conversation.