[avar]

I'm cautiously optimistic that increased AI use in the military is going to be a good thing for everyone.

Realistically we're not going to solve the problem of nation states wanting to protect their interests halfway across the world, which'll among other things mean killing some "combatants" from a drone.

But we can hope to do things like improve targeting, and a reduction in civilian casualties or collateral casualties. Right now the "AI" is some group of 20-somethings sitting behind a computer in Nevada, what if we trained an AI instead, and could e.g. hold legislative audits on what that software was configured to target?

[brokenmachine]

> I'm cautiously optimistic that increased AI use in the military is going to be a good thing for everyone.

> what if we trained an AI instead, and could e.g. hold legislative audits on what that software was configured to target?

You are much more optimistic than me.

If those audits ever happened, they would be held in secret, and have very different goals than most people would consider moral.