Robots vs Humans

http://bit.ly/bZFxB5

Robot arm punches human to obey Asimov's rules

I’m discussing this article because it has a close relation to a similar topic we watched a video on in class that I found interesting. A Slovenian researcher called Povse went against Asimov’s first rule of robotics, which states: “a robot may not harm a human”. In his experiments he specifically programmed a robot to deliver predetermined hits to a human subject at varying locations and speed.

 He asked for ethical approval from a Slovenian university and then persuaded six of his male colleagues to volunteer to help develop robots’ reaction time to the presence of humans and determine the parameters within which they can safely operate in our vicinty. Povse borrowed a Japanese robot that was assembled to make coffe vending machines. This robot was programmed to move to a point in mid air where it would meet a specific point on the volunteer’s body. This test was tried with each colleagues eighteen times with the robot primed to strike at varying areas on their person.

The robot’s tools varied between a round, blunt, and sharper object. The volunters then were asked to determine the pain level of the collision ranging between painless, moderate, and even unbearable pain. Povse himself tried it and concluded that the pain factor was between “mild to moderate range”. They will continue testing with more severe pain factors using an artificial human arm. This is in order to develop robots into sensing when a human is close and realise to stop moving in time before hurting the people around it. Povse presented his work in the Systems, Man and Cybernetics in a conference in Turkey.

 A worker named Sami Haddadin specialising in human and robot safety in the German Aerospace Centre announced that; "Determining the limits of pain during robot-human impacts this way will allow the design of robot motions that cannot exceed these limits." This same man earlier on the year put his own arms in the way of a knife-wielding kitchen robot to show how intelligent the robots with these sensors. Therefore he said it is crucial to develop these sensors even further if robots will ever work closely with humans. Even though this is true a a biomechanics specialist at Baylor College of Medicine in Houston called Liebschner argued that they shouldn’t test pain as it’s not essential to prevent pain but injuries.

 
I personally found this article very interesting since it didn’t only link with the videos we have been watching in the lectures but also because we can clearly see that our world is developing into more and more computerised machines. This means that the safer they are, the better the robots and humans can work together at the end.