War and military research throughout history have done many good things for quality of life, even back to the days of Archimedes fearing the Roman invasions, when it is more about defense not offense. This here ARPA/DARPA decentralized network of the internet is one of those good things along with self-driving cars, wifi, even microwave ovens come from radar research...and many others.
But what could possibly go wrong when military/war artificial intelligence is combined with all your data and knows everything about you? We have seen Terminator and heard about Skynet. What if what it knows about you is wrong or spoofed?
The ramifications could be much less than worst case Skynet but you can plainly see the way Youtube automation is even just for copyright, you gotta ask yourself, will it get everything right? Will YOU be deemed an opponent and a 'bad guy'? When the machine does it there is also a level of liability taken away.
Defensive military/war tools are one thing and lots of good from bad throughout history, but when defense is pre-emptive offensive actions like drones, the line gets fuzzy. Are we Archimedes or the Roman soldier that killed him?