more and more capable ie more and mora ton mora ton mouse. the computer watson is teaching itself. they have cableable of doing that. they learn faster than we can if they want. they work within very narrowly designed parameterparameters. we program them you must never harm a human being. it is the laws of robotics. that s what s going to protect us from anti social robots in the future. each one responds to the lab can t harm a human being you have to do what you are told, et cetera. unfortunately when we have already violated the first law in the way we design a lot of military systems. we want robots that are capable of carrying out complex military missions. so fine. does that mean they are going to pose a threat to the people that are operating them. not necessarily. it does mean this. as we ask them to do more and more things, as we design them