I found myself shouting at the radio again. (Chris even came out to my shop wondering if there were people in distress in our woods). I was listening to Science Friday (one of the best programs on NPR, the best network radio station in America)... and it just wasn’t good enough... again. There he was, a respected, highly intelligent and educated expert pontificating on the future of artificial intelligence and robots. I kept hearing his premise that AI’s and robots would be able to learn ethical behavior and empathy from human beings... I expect just a little push back from the interviewer when an a priori position of the guest has massive flaws.
I had a problem with this sort of thing way back in 2002, (see: http://somelightsomeserious.blogspot.com/2010/04/end-of-war.html ).
In this recent case the speaker seemed to be ignorant of a few obvious facts. 1). Presumably, most humans are born with emotions and at least some capability for empathy. These two things lead to compassion and hopefully, ethical behavior. AI’s are supposed to be able to learn these “higher functions” from us, their creators. HAHAHAHAHA. This gentleman was ignoring the fact that even given our innate (human) ability for emotion, empathy, the concepts of ethical behavior, and rational thought; we humans are slaughtering each other’s children all over the world. Which system of ethics? Which human beings? Empathy for which group of people? When do we start?
Asimov had it figured out 70 years ago. He knew that robots would have to have better ethics than we humans have. All these years later the experts don’t seem to have recognized the reality of their logic. IA and robots are unlikely to be any better than humans, their creators.