The Labyrinth: An AI Film Pipeline Experiment

Why We Did This

This is a project to better understand how AI impacts pre-production, treating it like an animated film and using Simon Stålenhag’s The Labyrinth as a loose reference.

The Labyrinth started as an experiment to better understand how current AI tools impact art direction and CG workflows - especially in development and pre-production, with an eye toward what holds up in production and post.

The Process

The project was treated like a full art department and world design pipeline, with some R&D extending into production and post - covering concept, environment and character creation, previs, shot layout, animation, lighting, rendering, and compositing - while deliberately comparing hand-built and AI-assisted approaches across each stage, using Unreal as the central hub, and using AI to explore multiple visual directions.

I led the artwork, visual and technical exploration, with Felix working alongside to frame the tests, critique the work, and dig into where; and why; the tools helped or didn’t.

The Lessons

It clarified where AI tools genuinely help, where they fall short, and how we might use them on future work.

We found a few areas where AI tools had real, practical value, and many more where they over-promised. What we walked away with was clearer, hands-on understanding of where these tools actually fit across the live-action, and digital art pipeline.

More importantly, it gave us a better sense of how we might use them differently; and more responsibly; on future projects.