Mathematical modelling consists of 3 components:

1. Assumptions: These are taken from our experience and intuition to be the basis of our thinking about a problem.
2. Model: This is the representation of our assumptions in a way that we can reason (i.e. as an equation or a simulation).
3. Data: This is what we measure and understand about the real world.

Current AI is strong on the model (step 2): the neural network model of pictures & words. But this is just one model of many, possibly infinite many, alternatives. It is one way of looking at the world.

In emphasising the model researchers have a strong implicit assumption: that their model doesn’t need assumptions.But all models do.