In reinforcement learning, the same approach (simulation) to addressing a data deficit is used, although the goal there is learning and not testing.
When certain events happen rarely, you have fake them deterministically so that you have reproducible coverage.
Unrelated to the original topic, but I had never come across that talk and it is great. I use the same basic approach to testing distributed systems (simulating all non-deterministic I/O operations) and that talk is a very good introduction to the principle.