Assume for argument’s sake that we have no idea how AI will transform the future.

This isn’t terribly unreasonable. People often forecast poorly when it comes to how, why, and when new technologies will be adopted. Maybe your first 5 thoughts are wrong. What are thoughts 6 through 10? What are thoughts 100 through 150? What is so out there that it is unrealistic? That might be the realistic option in 5, 10, 20 years.