The AI Guys Need to Work on Automating Politicians

This will be brief.  I'm reacting to Thomas Edsall's column today, which is more on the how robots have big positive effects on GDP, but big negative effects on the workers they displace.  We have yet to figure out a mechanism so all can internalize the benefit of automation.  One thought is that we aren't really trying hard to solve this problem, in part, because we have this bizarre and distorted notion of meritocracy.  So those who are being harmed by automation don't deserve any better.  What a myopic and ultimately foolhardy view of what is going on.

Some years ago I wrote a post called, The Economy as One Big Brain, where I deliberately tried to put the shoe on the other foot, so talked about some fiction which I called The Virtual CEO.  If we could automate the position at the top, might we start to focus on making more work for people down the line, as the virtual CEO would not require astronomic compensation to perform at a high level.  It is worth pondering along these line.

But a different thought should be entertained as well.  There is some presumption that AI works great with repetitive tasks, but is in over its head when the decisions require executive function and are therefore not nearly so routinized.  The question is this - do many people in executive positions nonetheless spend the vast majority of their time making routinized choices?

I don't know how to answer that question for CEOs.  But it seems to me the answer is yes for many politicians.  If so, we should start seeing The Virtual Congressman, if not The Virtual President.  I, for one would be very interested in an AI analysis of members of Congress, to see if the work could be automated or not.  My guess is that answer correlates with which party the member of Congress belongs, but I'd be interested in learning otherwise.