Man versus Learning Machine
Not a day goes by in my life when learning technology isn’t raised in a conversation. But how far can technology really go?
I read a very interesting article recently titled Where Computers Defeat Humans, and Where They Can’t by the New York Times. This set me thinking. What are the limits of where humans will be replaced by computers?
The article was about AlphaGo, an artificial intelligence system that has just beaten the human champion of the complex strategy game ‘Go‘.
The article explains how teaching computers to do complex human tasks such as playing a game of Go is extremely difficult. Much of the information and knowledge that humans store they can’t fully access to explain how it is that they are doing something, let alone then transfer this knowledge to a computer. This is a phenomenon that has come to be known as “Polanyi’s Paradox.”
Much more complicated than Chess, a game of ‘Go’ cannot be won by a computer through simple programming. AlphaGo’s enhanced success comes from the ability to not only learn winning strategies by looking into a huge library full of top strategies, but also from learnt information that had been reinforced by playing millions of games against itself.
It sounds to me that AlphaGo has been programmed much the way our brain works when it’s developing strategies for behavioural change, reviewing past successes and failures to make decisions about the next course of action. Of course, change is much more complex, and every individual is different so I can’t see a world where computers will have an algorithm prescribing habitual change just yet!
It’s so interesting to consider the implications of technology, not just for learning but on society as a whole. So we will continue to follow the man vs. machine dichotomy, but we must remember as commentators are saying that although in this scenario the machine is helping itself learn, it is still people who built and programmed it.