AI for president? (hear me out)
It wouldn’t have full power. It would just make some calls.
BUT, they wouldn’t be instant.
Maybe they only get applied 2 months after its decision or something.
And if something doesn’t seem good, the AI could be voted out any day, and everything it said stopped.
Then we wouldn’t need to vote every 4 years, but we should vote every now and then, to get the best AI in quick succession.
Start with smaller decisions, maybe budget tweak suggestions…
And if you’re thinking that it can’t be done by AI, what makes you think that it can be done by humans?
For example, it’s going to be a while until we can replace lawyers (if we can). But we can ALREADY use AI (trained on specific laws) to answer law questions, the answers are probably more factual than a lawyer, and you don’t have to pay 500$/h for it.
So why couldn’t you (eventually) train an AI on bureaucratic decisions.
We train on past decisions, and train the newer models on the answers from the public: “this AI was voted out because of decision X”.
I am NOT saying that we should do everything all at once, but that we could test the waters.
Kind of like they replaced some air traffic controller jobs with software.
At first, the software was running and giving warnings, and the controllers didn’t change anything about their job, it was just running in case they didn’t see something.
And the controllers always knew about what the software warned about, and then they gradually got the software better and better, until some parts of the controller’s jobs could be dispatched to that software. One at a time…