So, in order for AI to really be able to make decisions, we need smaller AI that people can run on their own laptops so they can test and verify that everything was done correctly in the big model. So, Bowling Green came out with this report and there are 100 policies that are recommended and you can sort them by the ratio of up votes to down votes and the number of votes which is a proxy for interest. You could imagine a next sorting step where you took a highly trained AI and you said, “Okay, assign a cost to each one of these,” and then you give it a budget and then you say, “Based on these preferences, what are the six to 10 things the Bowling Green government should fund next year.” The step after that would be to just automate it and say the default is that what the AI says we should do is what we’re going to do.

Keyboard shortcuts

j previous speech k next speech