The temptation will be to have computers, mainly, doing the monitoring and this could be a problem eventually.
I agree with you but as you point out later in this post these interim devices are really too dumb to incorporate any of these laws really and are also not likely to pose a threat or gain any functional consciousness.
That as well but I'm more worried about difficult to monitor communication between robots and advanced computers.
Agreed but more immediately it might already be appropriate to discuss international treaties on military applicatiions for autonomous machines. Maybe they should only be used for transport, reconnaisance and evac rather than direct fighting roles.