This is outstanding:
No, the Army isn’t turning tanks into robotic “killing machines,” as some excited headlines put it last week. Instead, as part of the military’s urgent push for artificial intelligence, an Army program called ATLAS is developing an AI that acts like a virtual soldier in the turret, one designed to assist the human crew.
Despite an alarming name, Advanced Targeting & Lethality Automated System, ATLAS is actually meant to help the humans spot threats they might have missed, prioritize potential targets, and even bring the gun to bear — but it will be physically incapable of pulling the trigger itself.
Perhaps I'm made of sterner stuff, but the name doesn't alarm me.
Helping gunners under armor identify targets is excellent. Indeed, if we do get a useful battlefield Internet working, ATLAS would be a great help to virtual infantry in the rear via reachback manning a remote weapon station on a vehicle at the forward edge the battle area.*
The idea that it is particularly dangerous to trust a soldier will not shoot a civilian after the computer identifies a target isn't completely insane--witness people who blindly follow their vehicle navigation system into obstacles.
But ATLAS will be far superior to the present situation that requires a soldier not to mistakenly identify and shoot a civilian. Our troops are actually pretty good at this compared to other countries (and I'm not even counting those who deliberately target civilians) and compared to the historical record of troops in combat. Training will, as always, be key to using the new technology.
So come on, getting help to identify and aim is a welcome addition to avoiding civilian casualties.
And if applied to enabling reachback for the squad, it would be a welcome addition to preserving the lives of our soldiers. Remember, just using a remote weapon system to take a soldier out of the turret doesn't help if the entire armored vehicle suffers a catastrophic hit.
And assuming ATLAS works, of course. It will be interesting to see whether gunners in training exercises just turn the damn thing off because of too many false positive targets identified--or if it fails to identify too many actual targets.
[*To be clear, that suggestion that I made in Infantry magazine at the link is not about getting rid of infantry. It is about preserving infantry assigned to infantry fighting vehicles while the infantry is moving with the tanks in their IFVs, and just a source of casualties if the IFV is hit by the enemy. The infantry would operate remote weapon stations on the IFV from the relative safety of the near rear. When dismounts are needed, they would be brought forward.]
