Published 23. May. 2016

Artificial Intelligence in Human Decisions

Internet of Things

by Garry Kasparov

header-915125_960_720

I was fascinated by the part of Professor Bostrom’s recent book, “Superintelligence,” about instilling human values into machines. As another of my current titles is chairman of the Human Rights Foundation, based in New York, this is an interesting disciplinary intersection for me! To take it a step further, do you see ways artificial intelligence and related technologies can impact the human rights of humans?

Our machines so far are agnostic, used for good or evil. They can connect the world to work, to play, or to form a terrorist network. The prospect of moral machine intelligence leads us to a tough question: are humans good examples for our machines? That is, a neural network learning from observed human behavior, what would it think? Is it possible to program empathy? How many decisions should be handed over to algorithms?

To take one simple example, self-driving cars are a big topic today, with a huge impact on nearly every industry and therefore a huge economic impact. It also creates interesting ethical implications broad and narrow, a global, real-world demonstration of the ethical thought experiment known as “The Trolley Problem.” Let us say that your car, connected on the network to every other car, has to decide whether to crash into another car in traffic, with a small chance of multiple fatalities or to swerve off a cliff, with a 100% chance of killing only you. Your property, your car, and its programmed understanding of the greater good of human life, playing an actuarial game in the real world.

Artificial intelligence will have an impact on every area of our life very soon and business leaders must examine not only how to use it like a tool, but how it will affect entire industries. This is not simply a matter of obsolescence, although that is a pressing concern, of course. We will always need humans making decisions. Humans establish the parameters, the priorities, and the time frames. So far, only humans can decide why something is important. Working together with robots and other AI will define our century, and anyone in a leadership position should be up to date on the limitations and potential of artificial intelligence.


Meet Garry Kasparov at the 600Minutes Executive IT event in Germany on the 5th of October 2016.