I have a morbid fascination with the idea of all-robot wars between countries with rough parity to each other (both countries have robot armies/navies/air forces). War always seemed sorta pointless to me but it becomes so much more so when you can quantify all the elements of what should make a military successful. If you know all the aspects of a robot force (power efficiency, sensor suite, decision-making, etc) and have a rough idea of the production capabilities of the warring countries then it seems like you could just crunch some numbers and have an idea of who would win in a war of attrition. Strategic objectives and planning might still be relegated to human generals, of course, but at that point the generals are practically just playing a real-time strategy game set in real life rather than in a computer.
Ah, but there's another layer: each side would work hard to keep the other side as uncertain as possible about its own capabilities, since that makes the other side less likely to make the optimal decision (eg, they might start a war they can't win, or fail to start a war even though they would win if they did)
Of course there's always uncertainty involved, but uncertainty of outcome doesn't really lend any more purpose to an armed conflict.
Is it going to be easier to start wars if it only means sending robots ? At least at the beginning before it escalates into full war.