Why breaking physics is the next challenge for AI animation

Written by Sophia East, Creative Lead at SLIK

AI is getting remarkably good at recreating the rules of our world. The next creative challenge is getting it to break them.

Generative models have strong learned expectations about how familiar things should look and behave. Paper bends and falls. Staircases are rigid and support weight. People expect stable footing when they walk.

That's useful when your idea follows those expectations. But what happens when it doesn't?

At SLIK, we've been exploring exactly that: using the right creative direction to push AI beyond its defaults, while keeping the result coherent and believable.

A recent project put this to the test.

The scenario sounds simple: a woman walks up a staircase made from sheets of paper.

For the model, those familiar elements create a series of contradictions.

The creative challenge of conflicting expectations

The staircase needed to look unmistakably like paper, but function like something solid enough to climb.

Our charcter needed to progress naturally upwards. Her feet needed to make convincing contact with each step. The paper could flex slightly beneath her weight, but couldn't suddenly behave like loose sheets. And she needed to walk with complete confidence because, within this world, nothing unusual was happening.

‍