【Hacker News搬运】自主武器：保护平民的虚假承诺（2022）

[hackernews]

Title: Autonomous Weapons: The False Promise of Civilian Protection (2022)

自主武器：保护平民的虚假承诺（2022）

Text:

hn link

Url: https://www.cigionline.org/articles/autonomous-weapons-the-false-promise-of-civilian-protection/

文章标题为《Autonomous Weapons: The False Promise of Civilian Protection》，作者是Branka Marijan。文章讨论了自主武器系统（AWS）的发展和潜在影响，特别是对平民的保护。作者指出，虽然一些国家认为AWS可以加强国际人道法（IHL）的实施并保护平民，但这种观点存在争议。文章提到，AWS的可靠性、脆弱性以及如何与载人平台互动等问题都尚未解决。此外，专家们对于如何将现有法律和规范应用于这些新武器也没有达成共识。

文章还指出，目前还没有证据表明AWS能够在复杂和快速变化的环境中，如战场，做出适当的响应。因此，所有关于AWS对平民和民用基础设施的直接和间接影响的声称都需要仔细审查。作者强调，需要严格的法规来确保在未来几十年内开发的系统真正保护平民和其他非战斗人员。这包括对武器系统的足够人类控制，以及对可以使用的自主武器系统的类型和使用的场景的限制。

总的来说，文章认为AWS的发展和部署存在重大风险，需要国际社会共同努力制定相应的法规和限制，以确保平民在武装冲突中的保护不受威胁。

Post by: YeGoblynQueenne

Comments:

wolverine876: There's another major issue with civilians and autonomous warfare machines of all sorts: Civilian control.<p>For all of history, military power depended most of all on population size (to provide soldiers) and economic resources (to equip them). Mad wizards were limited by population. However, mad wizards - megalomaniacal persons with lust for power and disregard of others - with enough resources might manufacture a machine army.<p>Also, note that these machines are often built for the US military by people who openly denegrate democracy and seem megalomaniacal; they have the technology and a lot of wealth. Could they build a factory somewhere and hide in their bunker until it's over?

wolverine876: 有；这是平民和各种自主战争机器的另一个主要问题：平民控制<p> 纵观历史，军事力量在很大程度上取决于人口规模（提供士兵）和经济资源（装备士兵）。疯狂的巫师受到人口的限制。然而，疯狂的巫师——对权力有欲望、无视他人的自大狂——拥有足够的资源，可能会制造出一支机器大军<p> 此外，请注意，这些机器通常是由那些公开破坏民主、看起来自大的人为美军建造的；他们拥有技术和大量财富。他们能在某个地方建一个工厂，躲在地堡里直到它消失吗；结束了吗？

beloch: The IDF has been using AI just to identify members of Hamas and watch for when they're home[1]. They then drop bombs on those homes, typically killing family and neighbors in the process. Humans still have oversight of this process. Humans drop the bombs. Acceptable kill ratios of up to 20 civilians to a single <i>junior</i> Hamas member have been deemed acceptable.<p>The notion that human oversight or involvement will protect civilian lives is no longer tenable. If AI merely chooses a target, there is no shortage of humans who will bend over backwards to justify dropping the bomb.<p>I'd have preferred not to bring the Israel-Hamas conflict into this, for obvious reasons, but it really is setting a precedent in the use of AI for war purposes. When both sides have dehumanized each other, there just won't be much pushback when a magic box says, "Kill". There's nothing unusual about Israeli's or Palestinians in this respect. I fully expect the same thing will happen in <i>any</i> war that uses AI in this way.<p>[1]<a href="https://news.ycombinator.com/item?id=39918245">https://news.ycombinator.com/item?id=39918245</a>

beloch: 以色列国防军一直在使用人工智能来识别哈马斯成员，并观察他们何时；我们在家[1]。然后，他们向这些房屋投掷炸弹，通常在此过程中杀死家人和邻居。人类仍然可以监督这一过程。人类投掷炸弹。可以接受的杀害比例是，一名初级哈马斯成员杀害20名平民，这被认为是可以接受的<p> 认为人的监督或参与将保护平民生命的观念已经站不住脚了。如果人工智能只是选择了一个目标，那么就不乏人类会竭尽全力为投下炸弹辩护<p> I-；出于明显的原因，我们宁愿不把以色列-哈马斯的冲突纳入其中，但这确实开创了将人工智能用于战争目的的先例。当双方都对彼此失去人性时；当一个魔盒说“；杀死&quot；。有；Israel；s或巴勒斯坦人。我完全预计在<I>任何以这种方式使用人工智能的</I>战争中也会发生同样的事情<p> [1]<a href=“https://；&#x2F；news.ycombinator.com&#x2F？id=39918245”>https://&#x2F；news.ycombinator.com&#x2F；项目id=39918245</a>

wolverine876: > Advanced militaries have clearly indicated their interest in such systems. Reasons range from speed of response and enhanced situational awareness to the ability to overwhelm the defence systems of adversaries.<p>My understanding of experts I've read over years:<p>Western militaries, at least, have for over a decade worked on keeping humans 'in the loop', for the obvious moral reasons and also for effectiveness - human-machine teaming is considered much more effective than either humans or machines alone.<p>Yet they are developing autonomous machines because they don't believe they can rely on communications. The measure-countermeasure contest between communication and jamming leans too much toward the latter and it becomes a serious vulnerability: The machine army could be entirely shut down by simply jamming communications.

wolverine876: &gt；先进的军队已经明确表示他们对这种系统感兴趣。原因包括反应速度和增强的态势感知能力，以及压倒对手防御系统的能力<p> 我对专家的理解；我多年来一直在阅读：<p>西方军队至少在十多年的时间里一直致力于让人类；在循环-中；，出于明显的道德原因和有效性，人机团队被认为比单独的人类或机器更有效<p> 然而，他们正在开发自主机器，因为他们没有；Don’我不相信他们能依靠沟通。通信与干扰之间的措施对抗竞争过于倾向于后者，成为一个严重的弱点：仅仅干扰通信就可能使机器军完全瘫痪。

ChrisMarshallNY: Whenever I hear "autonomous weapons," I think of this scene: <a href="https://www.youtube.com/watch?v=TYsulVXpgYg" rel="nofollow">https://www.youtube.com/watch?v=TYsulVXpgYg</a>

ChrisMarshallNY: 每当我听到“；自主武器；我想到了这样的场景：<a href=“https://；&#x2F；www.youtube.com&#x2F！watch？v=TYsulVXpgYg”rel=“nofollow”>https://&#x2F；www.youtube.com&#x2F；看v=TYsulVXpgYg</a>

pvillano: A landmine is an autonomous weapon. Landmines also kill a lot of civilians.

pvillano: 地雷是一种自主武器。地雷也造成许多平民死亡。