A new year means new tech trends. Forbes made a tech trends top ten. Every week one of these trends will be discussed. This week we cover the trend “artificial intelligence”, which seems to get smarter and smarter.
2017’s Tech trend 1: artificial intelligence
Human beings are intelligent species; even though we lose the board game Go –the most difficult board game in the world– from the computer now. In March, a computer, called AlphaGo defeated the best (human) player. AlphaGo is a self-learning algorithm; these kinds of computer programs are able to teach themselves some tasks and are in somewhat working as the human brain. Well, that is what we think. Because do we exactly know how the human brain works? And do we know how algorithms ‘think’? I think the answer to both of these questions is a simple no.
The victory by AlphaGo means an important breakthrough for Artificial Intelligence. Before this game, computers already taught themselves other games without any help. The progress of the past few years is outstanding; only four years ago there was great excitement about the algorithm of Google, which was able to recognize cats on its own. Today, that is only a piece of cake; a computer today is teaching itself to drive, to have conversations with people, to recognize objects on pictures, and to lip-read.
The ‘cats algorithm’ is free to use for everyone. One that uploads a private photo album to the servers of Google Photo does no longer need to search through thousands of photos before finding a specific photo; searching for a keyword is sufficient. Typing in ‘cat’ and all photos from your collection of a cat will be shown.
Change is coming
Sounds great, doesn’t it? But it is not a new concept. Back in 1951, computer-scientist Marvin Minsky built the first neural network. Since then, it has been coming to a halt for several times. But change is coming. The reason? Big players such as Google, Facebook, Uber, Amazon, Microsoft, IBM, and Apple see dollar signs. These high-tech companies buy out almost all the talent in this branch, which will start working in Silicon Valley. Huge amounts of money are invested in it, because the one with the best smart algorithms will attract the most clients. As a result, your smartphone understands what you are saying (“please dim the lights”) and will your car drive you, on its own, from A to B.
This rapid progress also brings questions, like where it will end and if humans will lose their autonomy to AI-systems. Reality? Well, the fact is that AI still is pretty dumb; while humans are able to do different tasks are most AI algorithms able to do only one thing, although, they are extremely good at it. They are so called ‘idiot savants’. How do we learn an algorithm two things? We don’t that. Yet. The second thing we don’t know how to do (yet) is to make two algorithms work together.
Another program is that smart algorithms work only properly after an overload of data. And it is that data –that is easily accessible and of which organisations have a lot– that is used for these algorithms. But what happens to the data, which might be even more important than the pictures of cats, but which is less accessible? 2017 might give us an answer.