July 11, 2017

Do we want to settle for working or middle class robots? I want the 1% top ones to work for my grandchildren

Sir, Sarah O’Connor while discussing the issue of jobs, for humans or robots, sensibly concludes that it is not “the routine jobs” taken over by robots that should bother us but “the basic stuff — homes, security, prospects — that we lost along the way” “The middle class is not shrinking as much as it thinks” July 11.

O’Connor brings up an interview from a 1974 book “Working” written by social historian Studs Terkel. In it a steelworker says: “I want my kid to be an effete snob . . . If you can’t improve yourself, you improve your posterity. Otherwise life isn’t worth nothing.”

I sure agree with this steelworker’s general concept, but, if my grandchildren must turn into effete snobs, I hope it is not because they have been replaced by some low or middle class robots, but by the 1% absolutely best ones… or the smartest ever artificial intelligence.

Sir, it should be clear that the better the robots that work for us the more they could produce for us. The marginal contribution of robots that substitutes for bank tellers must surely be less than that of robots that substitutes for bank CEOs.

Just as an example, let us suppose current bank regulations had been carried out not by Basel Committee technocrats, but by some smart artificial intelligence. Then the 2008 crisis and the ensuing slow growth would never have happened. Mr. AI would of course first have looked at what causes major bank crisis and so determine that excessive exposures to something ex ante perceived as risky, never ever did. He would also have understood that allowing banks to multiply with different leverages the net risk adjusted margins, would completely distort the allocation of bank credit to the real economy.

So what can we do? I would say first to make sure to keep the competitive pressure up on robot manufacturers. If we increase minimum wages for humans and do not begin taxing what the robots produce, we will not get the best robots we want.

An updated Chinese curse would be: “I wish your grandchildren live attended by 3rd class robots and dumb artificial intelligence.” And Sir, I would hate for that to happen to my grandchildren, because of something that I did or did not do.

Of course then we would come to the very delicate issue of how do we redistribute robot and automation productivity to humans. That is going to be awfully contentious. The only thing that occurs to me, before social cohesion breaks down, is to being by trying out a universal basic income.

That UBI should start out low and be very carefully designed. That is so because an UBI would become de-facto the robot that substitutes for the current redistribution profiteers, and so these would love to see it fail.