This is my take too. When we were imagining AI what were the use cases we had in mind back then? They are these grand visions of AI will take care of major problems. We should be pushing for responsible AI deployment, starting on low risk areas and moving up to more serious uses once we know the tools work for less catastrophic situations.