I remember reading a short story in an old anthology about a robot that has a limited memory that is filling up and will soon stop being able to make new memories. I think I remember that the robot had a weapon and is threatening a technician. The robot explains that it wants to control its own destiny, and kills itself.
The premise behind the story was that using a perceptron, an idea first proposed in the 50's, it would be simple to make an AI. The way this was done was to take a glass sphere, cover the inside with a magnetic coating, and write information on the inside. The underlining question was what to call the self-destruction of a AI. If it was suicide, did that imply person-hood?
The only reason I remember this any of this is I think it is the base of the robot shrink, Dr. Perceptron, in Futurama, yet I have can find no reference to this anywhere. Appropriately, it's been driving me nuts for years.