In the future, we might be creating simulations that contain artificial intelligences. These are going to be very difficult questions for regulators of the future to come up against. For example, is it ok to create a world with some suffering, as long as there's a lot more happiness? Or is creating any suffering going to be in itself a bad thing?