2015년 10월 18일 일요일

[Midterm Topic] What do we have to do before meeting an A.I. era?


It is time to come up with the moral code for Artificial Intelligence(A.I.). Everything around us is changing at a speed of light in a modern society. Just imagining what will come tomorrow is an exhilarating thing to do. Now we are facing an era of IoT(internet of things) and the robot. That means we will live convenient lives where the information on our every single activity is transmitted, saved and analyzed to ‘serve’ us in the best way. During this process, big data is generated. The designated and programmed machine (we can name it) will respond to the human’s need appropriately. How about the robot? It is in the beginning stage of development but its role is expanding in many areas from factories to hospitals replacing human workforce. Some robots even provide emotional service which is essential for healthy human life. Also, development of super intelligence(artificial intelligence) is on the process. If big data and  A.I. intergrated what would happen? A.I. is very similar to human intelligence, and very different at the same time. First, A.I. does not have a limit of volume. It could be built in a warehouse. So does its speed.  The last thing human needs to invent will be A.I. because after then A.l . will invent things much better than when humans do, but the content of the result matters. Suppose we give an A.I. the assignment to make humans smile. The A.I. could embed an electrode into humans’ cheek to cause a constant smile, because ,for A.I., the result of a task does not need to be worthwhile at the human’s standard. Now, Deciding the response to humans need from big data is in the hands of human. At the early stage of IoT. There do not seem to be many problems with it.


We all know a myth of King Midas. He wishes that everything he touches be turned into gold. He touches his daughter, she turns into gold. He touches his food, it turns into gold. This happens when we set misconceived or poorly specified goals. As such, the matter now is how to define human values, and load it onto the A.I. system, and IoT. A.I should be able to consider if the result accords with human values and to pursue it when performing action. The aspect  mentioned above should be put first to the development of both A.I and IoT before they become uncontrollable.  

reference : Ted.com 

Nick Bostrom: What happens when our computers get smarter than we are?

댓글 없음:

댓글 쓰기