Hacker News Reader

Comment by beloch | Hacker News Reader

[flagged]

What you're describing might be intended. Your argument hinges on the assumption that oversight was supposed to stop this, but it could be they just have a higher level of acceptable collateral damage.

Either way, it's the responsibility of humans to design risk controls. When the AI makes a mistake it should be auditable how it came to that decision. The humans need proper training and regular exercises and ultimately to be held accountable.

I believe this achievable for those that want to achieve it.

The system specifically target members when they're home with their families, even when theyhave chances to hit them during combat.

At this point it's just an excuse, they've found the best way to maximize the civilians they can kill but still use the "human shields" blood libel.

Whatever the rationale, it's set a baseline for "collateral damage".

And it has used AI to minimize collateral damage globally, but perhaps maximize acceptable collateral damage of a certain sort. It kills not just the target, but those who love them most... and who might feel the strongest sense of loss, otherwise.

I fear for this world, where any movement that challenges [illegitimately-held] power can die in a cradle of strategically bombed relationships. The Geneva Convention was not designed to grapple with this level of micro-targeting...

It is interesting/odd that this conflict of all going on at the moment brings forth the examples of AI utilization. Why not Russia-Ukraine, or Myanmar, or any other conflict?

What is the specific example of "AI" as opposed to "statistics" that is being used? It doesn't seem very AI to perform analysis of social graph to ID someone as a target and then tell your autonomous surveillance platform to watch that person's addresses.

https://en.wikipedia.org/wiki/List_of_ongoing_armed_conflict...

First, Gaza really is an ideal testing ground for this sort of AI targeting system. It's tiny. Pretty much all of it is under continual surveillance. Israel has completely modern precision bombing capabilities.

By comparison, just the front in the Ukraine-Russia war is massive. Many of the targets are nowhere near the front. The surveillance and precision air strike capabilities of both sides are inferior to those of the IDF.

Yet, we do see a lot of really savvy strikes being made on the basis of cell-phone data, etc.. This kind of intelligence could have been gathered by humans, but it's not impossible that either or both sides are already using AI targeting systems similar to Lavender. They're likely just not as effective for the reasons mentioned above, and others.

Militaries are likely not going to talk about this stuff much. Heck, this capability is in the same ballpark as things that past leaders may have sacrificed cities in order to hide[1]. Even the article on Lavender uses a lot of anonymous sources, so who really knows if it's true?

Second, it may not seem "very AI" to identify targets and determine when they're home. Yes, humans could do this. It would take a significant amount of time though. If the reports are true, Lavender is able to do this so quickly that Israel can bomb the homes not just of leaders and generals, but of raw recruits who've probably received the sum total preparation of a rifle and a one-time pep-talk. Quantity has a quality all it's own.

https://en.wikipedia.org/wiki/Coventry_Blitz#Coventry_and_Ul...

Reply on news.ycombinator.com