is a strange thing. You cannot give real autonomy to a machine, you can only make it mimic autonomy, but in any event you are only overlaying your own desires onto a machine.

The Mars rover “Curiosity” acts autonomously- the distance between the rover and Earth is too great to allow fast enough communications so that it can be operated remotely- but what it does have is a set of instructions that operate within a limited set of parameters. Go places- move around. Avoid big obstacles and stay out of shadow. Pick up dirt and test it. And as simple as these tasks are, it has a pretty substantial program to do them.

Autonomy requires thinking for ones self. And not just thinking for one’s self, but making choices and decisions not based on your programming, but on independant choices you make based on your environment.

Here’s an example. A robot sees a bird. It’s programming lets it recognize it as a bird. It looks up “Bird” and decides what to do with the bird. Should it catch it and make a photographic record of it’s activities? Should it try to dissect it and see if it is healthy, or deformed, or similar to other birds? Or is it’s bird study already done, and the robot should ignore it?

All of those decisions are made based on the program of another person, not any autonomy on the part of the robot. The robots actions only mimic autonomy, but if the program is known, the actions are completely predictable

A human, with true autonomy, sees the bird and wishes he could fly. Imagination is the key to autonomy, the idea of seeing something and imagining things that have never existed before and attempting to make them happen. Even if you could write code that “Mimics” imagination, it is still only a poor substitute. You could write a program that allowed a robot to hear a song and then decide it wants tolearn to play the piano, but you can’t make it start writing songs on it’s own.