One wonders if weaponized artificial intelligence won’t someday short-circuit itself into attacking human civil society. It could happen unless aritificial inteligence is always made with a law degree and legal sensibility concerning the rules of war (joking). Artificial intelligence made to command and control ordinance to kill human combatants, and they are difficult to identify sometimes in irregular wars of today, may be instructed or find their own way to kill the innocent (well, if they voted for Hillary perhaps not so innocent).
Weapons are too easy to invent. If people were so creative with ecological zoning and tech to conserve habitat for wildlife and humans it would be better. There are innumerable ways to air-drop anti-tank munitions with A.I. able to attack tanks, and to camouflage them. i.e. They could look like dead animals then produce a rocket as a tank becomes visible in its cold dead eyes. Revenge of the road killed. Dissolving parachutes could place hover-drone warheads disguised as cow pies. etc. ad nauseum. One must stipulate that the video is better than Star Wars; Attack of the Clones.
While the A.I. in the anti-tank systems can be pretty stupid, weapons designers aren’t likely to be content with stupid artificial intelligence systems and inevitably will want their A.I. creations to be smarter than the average and much smarter than the enemy artificial intelligence. The United States has built a super-computer that is the biggest in the world; more than twice the size of the Chinese former world record holder, yet is that smart enough? Is it used a little bit for out-thinking the enemy in tactical situations (hen winning is peace and prosperity all around with a recovering ecosphere?)
Will artificial intelligence military systems given enough intelligence to make a talking rabbit able to trick humans into believing it voted for Trump, will the A.I. become some sort of peace nik with the belief that war is stupid?