Hacker News new | ask | show | jobs
by chrisco255 2715 days ago
Isn't it inevitable that AI will be used to improve targeting technology for weaponry? Am I wrong in assuming that if U.S. doesn't develop this tech some other country will?
5 comments

It’s not inevitable if we aren’t willing to let it happen. It used to be “inevitable” that poison gas, napalm, cluster mines, etc. were inevitably the future of war, but as a species we decided we weren’t okay with the effects of those technologies, and we’ve subsequently been reasonably successful at not using them.
I mean, all of those things are still being used, though....
Don't let the perfect stand in the way of the good. If the treaties help prevent 90% of the usage, they seem worth having.
That doesn't work in the real world where the remaining 10% of countries are willing to use their superior military technology to dominate others.

The only way for a country to be truly free is to be either equal in power to other countries - or barring that, the most powerful country on the block. Since the former is impossible due to the nature of reality, all countries aim for the latter.

In practice that top 10% is generally the one holding itself accountable to arms treaties, it's normally smaller groups that go for the cheap-and-horrible.
Imagine if all engineers had boycotted weapons development due to the heinous nature of napalm firebombing in WW2. US military competency stagnates at the stage of napalm. Would that actually change the behavior of the government's military, or would they continue to use napalm in Vietnam and beyond? As history has borne out, I think the latter is the case. The gov will assume it is the best tool so far to do the job, and preferable to the prior alternative of mass cannon bombardment and infantry deployment.

Development of smart munitions, better sensors/intel, and targeting precision has reduced the scale of military operations, entrenchment, and collateral damage. I think that was a form of technological disruption that was overall for the better.

There's a valid counter-argument that making war smaller and easier will lubricate the willingness for politicians (and the public) to enter into war, or maintain a state of pseudo-war. That is certainly a drawback.

This is the arms manufacture argument. Sure, smarter, more efficient killing tools may seem like a benefit, but it's always framed in a "us vs. them" argument.

What happens when you're the "them" at the receiving end of these smart weapons? Weapons tech is a pandora's box, once opened, everyone has it and you can't close it.

The progression was always towards more targeted weapons, because they have less collateral damage.
And a much higher chance of actually destroying the target. The US dropped more tonnage of bombs on Vietnam than the entirety of WWII, and yet random destruction is rather ineffective.
> Am I wrong in assuming that if U.S. doesn't develop this tech some other country will?

You are not wrong. Alibaba, Baidu, etc. work heavily with the Chinese government in this area.

Is that the old tired "if you won't help build tools to butcher people, evil chinese will?" argument?

This zero-sum, jingoisti, outlook on the world has caused most damaging and bloody wars in history. World policy is not a zero sum game and every human does not have to help kill other humans for the world to be at peace.

> "if you won't help build tools to butcher people, evil chinese will?"

It's not that they will. It's that they are.

When someone who most certainly does not have your best interests at heart is building technology capable of crippling or dominating you, what response do you suggest? Passivity?

The only recourse here is to either convince them to step down or to - at the very least - match them. Anything else puts you at a disadvantage and puts your citizens in danger.

> It's not that they will. It's that they are.

[Citation needed]

> When someone who most certainly does not have your best interests at heart is building technology capable of crippling or dominating you, what response do you suggest? Passivity? > The only recourse here is to either convince them to step down or to - at the very least - match them. Anything else puts you at a disadvantage and puts your citizens in danger.

You are presenting a false fallacy where the only two options are "do nothing" and "invest all possible resources available to the country into killing other people". A casual look at history book would quickly show you that escalating arms races are not beneficial in the long run - even to the country winning them, since they degenerate into building murder tools at the expense of their own citizens. Much like modern US, which is incapable of providing healthcare to their citizens.

That does not mean that no resources should be put into defense, but every civilian IS NOT morally obligated to help kill other people.

Do you really need a citation for what the Chinese government is up to with classified military technology? (As if the Chinese government willingly publishes this info)

I would argue that the very raw destructive power that military weaponry has, in particular with regards to nuclear capabilities, has actually reduced the odds of another world war.

> [Citation needed]

The Chinese government has imprisoned 1 million+ members of a ethnic minority. I'm sure you've heard of this, if not, you need to do some research...

> You are presenting a false fallacy where the only two options are "do nothing" and "invest all possible resources available to the country into killing other people".

No one is presenting that.

> A casual look at history book would quickly show you that escalating arms races are not beneficial in the long run

[Citation needed]

From my perspective, it sure did help, as nuclear weapons resulted in the inability for large countries to go to physical wars (for now, at least).

> World policy is not a zero sum game

The Pentagon does not deal in World Policy, it deals in martial policy. As General Mattis said, "If you don't fund the State Department fully, then I need to buy more ammunition ultimately".

And as we've seen, it'll eat all the money you can throw at it and even more, even if it means no healthcare, education or basic infrastructure in the country they're supposed to defend. Militarys appetite for money is limitless.
> and even more

I find it hard to believe the military spends more than Congress gives it. Military doesn't decide the budget either. Of course a lot of pork is reps spending money in their districts to "create jobs", even when the military doesn't want what they're making [0]. Clearly Pentagon isn't in charge of that spending, or they wouldn't be spending money on what they don't want.

[0] https://www.military.com/daily-news/2014/12/18/congress-agai...

>Military doesn't decide the budget either.

This relies on the (wrong) assumption that military leadership doesn't hold sway in the government. Historically, when the pentagon asks for more funds, the USG provides.

No, there is a difference: if you aid the US or not, the US will butcher their enemies. Whether or not you aid the US, companies are helping China butcher their enemies... but who are China's enemies? The US is one of them.
Face recognition is categorically inevitable for small targeting systems.

The danger with AI in particular is that eventually the technology gets trivially accessible. You don't need to hire specialists, just wait a bit and download the library and get the how-to book from Amazon or Barnes and Noble.

That’s a rather common argument. The glib answer is a reference to the Eichmann trial. To (slightly) elaborate, in such a scenario it would at least be only the second (or theirs, fourth...) best team working on the technology.

There are also instances of successfully preventing technology from becoming commonly used in warfare. Nuclear tech is one, but not a perfect fit because it requies huge investments. But chemical weapons are quite comparable to AI, in that any chemistry grad could create such weapons, yet they have been used far less often than mere feasibility would suggest.

If AI can reduce friendly-fire or weapons striking the wrong target, isn't that saving lives where both of those examples currently kill an extraordinary number of innocent people?