Augmented Human Intelligence

One of the leading proponents of man-machine integration is Gary Kasparov.
It is ironic, given that his defeat by Big Blue was one of the landmark moments of artificial intelligence progress. But he managed to turn the tables on his opponent beyond the chessboard and push for the idea of Freestyle Chess, where you are allowed to compete on human only, machine only, or mixed teams. In these competitions we are able to see that both man and machine only teams are at a disadvantage when faced with the mixed teams. This goes to show that neither side has an absolute advantage, neither side is truly superior, and we can Augment Human Intelligence.  (A.H.I.?)

What are the practical implications of taking this view point?  How would it change the way we relate with the machines? How does it change the future of work in the 21st century? What about art? Science?

How could AI help traders?

 
We could have a.i. help traders keep an eye on a much broader spectrum of instruments. Where the tasks of scanning the market for specific conditions is automated, a trader is presented with an opportunity containing relevant risk metrics and a recommended exposure based on Kelly criterion. From there the trader can make a decision on how to trade this idea, buy and hold? scalp around the idea buying the dips and selling the rallies? a derivatives play? Once the decision is made, the execution of the trade is also automated with the machine carrying out the necessary steps to accumulate the position and later distribute it for taking a profit or stopping a loss. 

At first, these tools wouldn't be very different from today's technical analysis indicators and order execution algorithms. But they could evolve by taking into consideration each trader's personal style, level of sophistication and current capital allocations.

towards augmented intelligence

I think we're going to see a shift towards "Augmented Intelligence" where the machine works more like a tool than as a replacement for people. In a symbiotic process. As bio-tech, virtual reality and nerve-machine interfaces evolve, this symbiotic relation will become much tighter. We are going to see cyborgs way before we see artificial general intelligence, of the kind you see in terminator movies.
  
With all the advances in machine learning, neural nets, deep learning, in the last few years. We've seen machines become really good at tasks where we can infer from millions of examples. But they aren't good at tasks which require decisions based on a unique set of conditions, and where every instance is unique and must be considered on its own.  Us humans are quite the opposite, we are great at judging unique situations, but really bad at inferring from large populations. Statistical thinking doesn't come naturally to us. I believe that the easiest way to bridge this gap is by having machine and person work together, each playing to its strengths. 


What do you think?