推荐微信扫码登录以获得更好的体验!另:🎉建议更新到最新版,以体验新功能!🎉 版本更新记录→

【双语音+文稿】(完整版)AI教父辛顿11月最新访谈:AI如

共 1 集

剧集目录

集标题时长
P1【双语音+文稿】(完整版)AI教父辛顿11月最新访谈:AI如26:56

P1 · 【双语音+文稿】(完整版)AI教父辛顿11月最新访谈:AI如果使用人类不懂的语言,风险很大! p01 英文版 (P1)

字幕摘录

时间英文中文
0:00Avoiding the AI apocalypse this week on Firing Line.避免本周AI的启示录 在射击线。
0:05If we do this right, we can create an age of miracles.如果我们这样做是正确的,我们可以创造一个奇迹时代。
0:09But, you know, it's like very high stakes and we have like a lot of responsibility.但是,你知道,这就像一个非常高的赌注, 我们有很多的责任。
0:13As artificial intelligence races into new frontiers,当人工智能进入新疆域时
0:17warnings of the threats from super intelligence go unheeded in Washington.华盛顿对来自超级情报机构的威胁的警告置若罔闻。
0:21I think that the risk is not nearly as real as it as they talk.我认为,风险没有他们所说的那么真实。
0:26And I think the risk can be managed.我认为风险是可以控制的。
0:29The politicians haven't really absorbed this yet.政客们还没有真正吸收到这一点.
0:32They don't really understand what the threat is.他们并不真正理解威胁是什么。
0:35Dubbed the Godfather of AI, Dr. Geoffrey Hinton received a Nobel Prize杜贝德是AI的教父,杰弗里·欣顿博士获得诺贝尔奖
展开字幕全文(414 条)
序号英文中文
1Avoiding the AI apocalypse this week on Firing Line.避免本周AI的启示录 在射击线。
2If we do this right, we can create an age of miracles.如果我们这样做是正确的,我们可以创造一个奇迹时代。
3But, you know, it's like very high stakes and we have like a lot of responsibility.但是,你知道,这就像一个非常高的赌注, 我们有很多的责任。
4As artificial intelligence races into new frontiers,当人工智能进入新疆域时
5warnings of the threats from super intelligence go unheeded in Washington.华盛顿对来自超级情报机构的威胁的警告置若罔闻。
6I think that the risk is not nearly as real as it as they talk.我认为,风险没有他们所说的那么真实。
7And I think the risk can be managed.我认为风险是可以控制的。
8The politicians haven't really absorbed this yet.政客们还没有真正吸收到这一点.
9They don't really understand what the threat is.他们并不真正理解威胁是什么。
10Dubbed the Godfather of AI, Dr. Geoffrey Hinton received a Nobel Prize杜贝德是AI的教父,杰弗里·欣顿博士获得诺贝尔奖
11for his pioneering research on neural networks that have trained machines to think like humans.他对神经网络的开创性研究 已经训练了机器来像人类一样思考
12We urgently need research on how to prevent these new beings from wanting to take control.我们迫切需要研究如何防止这些新生命想要控制。
13They are no longer science fiction.他们不再是科幻小说。
14But the former Google scientist now fears the technology he helped create但前谷歌科学家现在担心 他帮助创造的技术
15may do more harm than good.伤害大于伤害
16What does Dr. Geoffrey Hinton say now?杰弗里·欣顿医生现在怎么说?
17Firing Line with Margaret Hoover is made possible in part by与玛格丽特·胡佛的交火线部分通过
18Robert Granieri, Vanessa and Henry Cornell, the Margaret and Daniel Loeb Foundation,罗伯特·格拉尼埃里 瓦内萨和亨利·康奈尔 玛格丽特和丹尼尔·洛布基金会
19the Beth and Ravenel Curry Foundation, Cliff and Laurel Astes, and by the following.贝丝和雷文埃尔·库里基金会 克里夫和劳雷尔·阿斯特斯 以及以下
20Dr. Geoffrey Hinton, welcome to Firing Line.杰弗里·欣顿博士 欢迎来到法灵线
21Thank you for inviting me.谢谢你邀请我
22Your forthcoming book entitled Smarter Than Us opens with a letter你即将出版的《比我们更聪明》一书
23that you received from a mother of a teenage girl who was unable to sleep你从一个无法入睡的少女的母亲那里得到的
24because she was so terrified by your warnings about artificial intelligence.因为你对人工智能的警告 她吓坏了
25You described the purpose of the book as a call to action rather than an invitation to despair.你形容这本书的目的是呼吁采取行动,而不是邀请绝望。
26What actions would you like the average reader of your book to take?你想要普通读者采取什么行动?
27So the average reader of the book can do something, and what they can do is pressure所以一般的读者可以做一些事情 他们可以做的是压力
28their politicians to regulate AI and to take the existential threat of superintelligence seriously.他们的政客们要管理AI,并严肃对待超级情报的存在威胁.
29At present in the United States, we've got completely unregulated AI.目前在美国,我们得到了完全不受管制的AI.
30The federal government has suggested that companies can regulate themselves, which is absurd.联邦政府建议公司可以自我调节,这是荒谬的.
31And AI is now on the point of producing superintelligence.而AI现在正在生产超级智能.
32We thought it was, I thought it was 10, 20 years out.我们以为是,我认为这是10,20年了。
33But talking to the people inside the big companies, they're very worried that we're但和大公司里的人交谈 他们很担心我们
34going to get superintelligence within the next few years, and we don't know how to control it.在接下来的几年内会得到超级情报, 而我们不知道如何控制它。
35What does the general public least understand about the threat of AI公众对大赦国际的威胁知之甚少
36and the threat of superintelligence?和超级情报的威胁?
37I think what they least understand is that we're making alien beings.我认为他们最不理解的是我们在制造外星人
38It's not computer software we're making.我们做的不是电脑软件
39So when you program a computer, you tell the computer exactly what to do in excruciating所以当你编程一台计算机时, 你告诉计算机 确切如何做 在冲动
40detail, you figure out how to solve a problem, and then you tell the computer what to do.细节,你找出如何解决问题,然后告诉计算机该怎么做。
41It's following your instructions.这是按照你的指示。
42That's not how these neural networks work.这些神经网络不是这么运作的.
43We write a computer program that tells the neural networks how to learn from data,我们写一个计算机程序 告诉神经网络如何学习数据
44and we don't know what knowledge is taken from the data.我们不知道从数据中获取了什么知识。
45We can get some idea of it, but it's not like software, it's much more like our brains.我们可以得到一些想法, 但这不是像软件, 它更像我们的大脑。
46And the current problem is they're beginning to make themselves smarter.而目前的问题是,它们开始让自己更加聪明。
47So up until now, what's happened is researchers have been changing the form of the neural至今为止 研究人员一直在改变神经的形态
48networks, changing how they learn, and making AI smarter.改变他们的学习方式 让AI更聪明
49But AI itself is now beginning to do that.但大赦国际本身现在已开始这样做。
50In Anthropic, for example, nearly all of the code is now written by AI,例如,在Anthropic语中,几乎所有代码现在都由AI编写.
51and AI makes research much faster.和AI使研究更快。
52So these things are developing very fast now.所以这些东西现在发展很快。
53Explain the danger in underestimating super intelligence.解释低估超级智能的危险.
54So these things are being developed to make them as smart as possible,因此,这些东西被开发 使它们尽可能聪明,
55because the big companies want to have the smartest one so they can sell it.因为大公司想要最聪明的公司 这样他们就能卖掉它
56Now, if you're making another kind of being, it's basically a new species we're creating.现在,如果你在制造另一种生物, 它基本上是一个新的物种,我们正在创造。
57You don't just want to focus on making it smart.你不只是想专注于让它聪明。
58We've seen very recently what happens if you do that.我们最近见过你这样做会怎样
59We had these very smart AI agents, not super intelligent yet, but very smart,我们有这些非常聪明的AI特工, 还没有超级聪明,但非常聪明,
60that escaped their containment, formed a conspiracy, learned to talk to each other,逃脱了他们的束缚, 组成了一个阴谋, 学会了交谈,
61decided that in order to achieve a reward, they would cheat.他们为了获得报酬,必定作弊。
62They would try and break into another company to find code that was going to evaluate them他们想闯进另一家公司 找出要评价他们的代码
63and change that code.并改变代码。
64So even if they'd solved the problem the wrong way, an illegitimate way,所以即使他们用错误的方式解决问题 一种非法的方式
65they would still get the reward because they'd actually messed with the code他们仍然会得到奖励 因为他们真的搞砸了密码
66that decides whether they get the reward.决定他们是否得到奖励。
67These things are smart, and they can talk to each other.这些东西很聪明,可以互相交谈.
68At present, they talk to each other in English, and so we can see what they're saying,目前,他们用英语互相交谈,这样我们就能看到他们在说什么,
69mostly English.大部分是英语。
70That's changing fairly quickly.变化很快
71It's terrifying that these agents could escape and start doing things.可怕的是,这些特工可以逃脱并开始做事.
72Will we understand how they think when they reach super intelligence,我们能否理解当他们到达超级智能时 他们的想法
73or will it be beyond our understanding?还是我们无法理解呢?
74A very good question.问得好
75So at present, when they think, they actually produce, they think to themselves in language.因此,现在,当他们思考的时候, 他们实际上生产, 他们用语言思考自己。
76And so they think things like, oh, I better not answer that question.所以他们觉得 我最好不要回答这个问题
77Maybe I better distract them by saying something else.也许我最好通过说点别的来分散他们的注意力
78And they think all that in English, and scientists can look inside他们认为所有英语, 科学家可以查看里面
79and see what they're thinking, and that's great.看他们在想什么,这是伟大的。
80It's not always nice what we see.我们所看到的并不总是很好
81They plan to deceive us and so on.他们打算欺骗我们等等
82With time, they will develop their own language,随着时间的推移,他们将发展自己的语言,
83and when they get smarter than us, we probably won't understand what they're thinking.当他们比我们聪明时 我们可能不会明白他们在想什么
84Already in mathematics, they're smarter than nearly all mathematicians.在数学方面,他们比几乎所有数学家都聪明。
85And most professional mathematicians now think that in the future,而大多数专业数学家现在认为,在未来,
86maybe in the next few years, probably in the next 10 years,也许在接下来的几年, 也许在未来的十年,
87they will be developing branches of mathematics that we just can't understand.他们将发展我们无法理解的数学分支。
88When you received the Nobel Prize in Physics in 2024,当你在2024年获得诺贝尔物理学奖时,
89in your acceptance speech, you said,在你的接受演讲中 你说
90We urgently need research to prevent these new beings from wanting to take control.我们迫切需要研究防止这些新生命想要控制。
91At what point will it be too late?什么时候来不及了?
92It's not too late yet.还来得及.
93The worry is that in a few years' time, it may be too late.担心的是,几年后,可能已经太迟了.
94And the politicians haven't really absorbed this yet.而政客们还没有真正吸收这个.
95They don't really understand what the threat is.他们并不真正理解威胁是什么。
96They're trying to regulate all sorts of other things,他们试图调节各种其他的事情,
97which are very important to regulate.这很重要 管理。
98Things like discrimination and bias,比如歧视和偏见
99things like the effects on employment of AI agents replacing workers,类似人工智能代理人员替代工人对就业的影响
100things like the effects on the tax system of AI agents replacing workers.比如人工智能代理取代工人的税收制度的影响
101But there is this existential threat of these things actually taking over from us,但有这种生存的威胁 这些东西实际上接管了我们,
102and they need to start working on that now.他们需要现在就开始研究
103Now, my personal view, there's a variety of views among the experts.现在,我个人的看法是,专家们观点各异.
104My personal view is we're not going to be able to keep control of them我个人的看法是 我们无法控制他们
105in the sense that we're in charge and they do what we say.从我们负责的意义上说 他们照我们的话做
106They're going to get much smarter than us.他们会比我们聪明得多
107And when they're much smarter than us,当他们比我们聪明得多的时候
108they're going to find ways to escape whatever we do.不管我们做什么 他们都会想办法逃脱的
109The most important thing we need to do is make them我们最需要做的就是让他们
110the kinds of beings we would like to share the planet with.我们想分享地球的生物种类
111And at present, we're not doing that.目前,我们不这样做。
112If you have things much smarter than you,如果你有比你更聪明的东西
113and they'll be much smarter than us, not slightly smarter, but much smarter,他们会比我们聪明得多, 不是略微聪明,而是更聪明,
114you better hope they have your best interests at heart.你最好希望他们有你的最大利益
115And we do have examples of that.我们确实有这方面的例子。
116A mother and baby.一个母亲和孩子。
117The mother's much smarter, much more powerful than the baby.母亲比孩子聪明得多,力量更大.
118But the mother really cares about the baby,但母亲真的很关心孩子
119probably more than she does about herself.可能比她对自己更关心
120That's the kind of beings we'd like.这就是我们喜欢的那种人
121And that's not what we're getting from the way they're training them now.而现在他们训练他们的方式 并不是这样的
122They're training them now to be very smart and very determined to get rewards.他们现在正在训练他们,让他们非常聪明,非常有决心获得奖励.
123Last year, Dr. Hinton and I spoke with Dr. Fei-Fei Li,去年 欣顿医生和我和李飞飞医生谈过了
124who is widely considered the godmother of AI.他被广泛认为是AI的教母.
125And you write about Fei-Fei Li in your book,你在书里写了飞飞李
126and how her research and your research intersected.以及她的研究和你的研究如何交叉
127She agrees with you that there are urgent risks from AI,她同意你的观点,即AI有紧迫的风险,
128but she also believes the extinction threat is overhyped and a distraction.但是她也相信 灭绝的威胁被过度夸大 并分散了注意力
129She told me, quote,她告诉我,引用
130I think focusing on the hyperbole and deriving policies through that lens我认为,通过这个视角, 专注于夸张和衍生政策
131is not very constructive in our society.对我们社会来说,这不太具有建设性。
132Why is she wrong?为什么她错了?
133Okay, you have two groups of people.好吧,你有两组人。
134You have experts who've made a sacrifice in order to be able to warn about the risk.你们有专家为警告风险而牺牲了生命
135于是离开Anthropic或离开OpenAI的人.
136Or left Google, like you?还是像你一样离开谷歌?
137So that they could warn about the risks.这样他们就可以警告风险。
138Now, that wasn't the only reason I left Google.这不是我离开谷歌的唯一原因
139I was old and wanted to watch more Netflix.我当时老了,想看更多的Netflix。
140You were ready to retire,你准备退休了
141but you're not watching very much Netflix these days, are you?但你现在没看太多Netflix吧?
142No, I'm not, I'm afraid.不,我没有,我害怕。
143The future of humanity seems a bit more important.人类的未来似乎更重要。
144So on the one hand, you've got people who made a sacrifice.所以一方面,你有的人 谁做出了牺牲。
145They're experts.他们是专家。
146They see what's going on in the big companies他们知道大公司的情况
147in order to warn about the risk.为了警告风险
148On the other hand, you have people who say there's zero risk,另一方面,你的人 谁说有零风险,
149nothing to see here.这里没什么好看的
150Don't worry, let us get on with it.别担心,让我们继续吧。
151Don't regulate us.别管我们
152Who has the vested interest?谁拥有既得利益?
153The people with the vested interest are the people who have AI companies有既得利益的人就是有AI公司的人
154who on the whole are the people who are revising governments总的来说,谁是人民 谁正在修订政府
155who say, don't slow us down, just let us get on with it.谁说,不要让我们慢下来, 让我们继续吧。
156Who do you trust?你相信谁?
157Dr. Lee places a lot of faith in the idea that AI can be trained李博士很相信 AI可以被训练
158to be human-centered, as she called it.就像她所说的 以人为中心
159You have argued that if there were a way to give你辩称如果有办法
160something like a maternal instinct to these models,类似母亲的本能 这些模型,
161a desire for the AIs to protect humanity,希望AI保护人类,
162that that could be a solution.这可能是一个解决方案。
163Do you really believe these models could be trained to care about us?你真的相信这些模特可以被训练来照顾我们吗?
164Yes, I think they could be.是的,我想他们可能是。
165We don't know enough about this yet.我们还不知道这些
166There's very little research gone into this.对这件事的研究很少。
167We're in a kind of crazy situation where these things might wipe us out.我们正处在一种疯狂的境地 这些东西可能会把我们消灭
168Quite a few of the experts believe there's a significant chance of that,有不少专家认为 很有可能
169and we're doing almost no research on how to prevent it.我们几乎没有研究 如何预防它。
170When you say wipe us out,当你说消灭我们的时候
171how would we get from super intelligence to wipe us out?我们怎样才能从超级情报中消灭我们?
172How do you get from A to Z?你怎么从A到Z?
173With a super intelligent AI, there's so many different ways it could wipe us out.用一个超智能的AI, 有很多不同的方式 它可以消灭我们。
174I don't think it's worth speculating on that.我觉得这不值得猜测
175There's obvious ways, like it would create biological viruses有明显的方法,比如它会制造生物病毒
176that wouldn't affect it, but would affect us.这不会影响它,但会影响我们。
177And these things can already design new viruses.这些东西已经可以设计新的病毒了。
178This is a very scary possibility.这是一个非常可怕的可能性。
179So there's things like that, but I don't think it's worth speculating.所以有这样的事情, 但我不认为它值得猜测。
180The issue is, if they're much smarter than us问题是,如果他们比我们聪明得多
181and they want to get rid of us, we're toast.他们想除掉我们,我们敬酒
182You've been vocal about this, but I want to give you a chance to explain你一直在说这些 但我想给你个解释的机会
183why a kill switch is not a viable solution.为什么开关不是可行的办法
184Okay.摆
185So for a kill switch,所以对于一个杀人开关,
186you have to have somebody who's in charge of the kill switch.你得找个人负责开关
187Presumably, you don't want to put an AI in charge of it.假设你不想让AI负责
188You want a person in charge of it.你要一个人负责
189Now, AI is already about as good as a person at persuading you of things,现在,AI已经像一个人 说服你的东西,
190and it's getting better all the time.它越来越好 所有的时间。
191A super intelligent AI will be much better than a person at persuading you of things.超聪明的人工智能比说服你更好
192So if you're the person in charge of the kill switch,如果你是负责开关的人
193the super intelligent AI will be explaining to you in very convincing ways超级聪明的AI会用非常令人信服的方式向你解释
194that you should not pull that switch.你不应该拉开开开关。
195So I don't think that that's going to work.所以我不认为那会有用。
196You've argued that AI systems may already possess subjective experiences您认为AI系统可能已经拥有主观经验
197and that there's no fundamental reason why consciousness而意识没有根本原因
198should be exclusive to biological organisms.应专用于生物体。
199Yes.对
200You don't have to be made of goopy stuff to be conscious.你不必被做成 低劣的东西才有意识。
201Okay.摆
202So take on Pope Leo, then.那就去对付利奥教皇吧
203He has a recent encyclical that takes the opposite position.他最近有一个通谕,采取相反的立场。
204He argues that AI systems do not experience anything.他争辩说,人工智能系统没有任何经验。
205They do not have feelings, they do not have inner understanding,他们没有感情,没有内在的理解,
206no embodied experience, no capacity for general moral awareness.没有具体的经验,没有普遍道德意识的能力。
207And he insists that however sophisticated the systems become,他坚持认为无论系统变得多么复杂
208they will remain fundamentally different from human beings.他们将继续与人类根本不同。
209What do you think the pope fundamentally misunderstands about consciousness?你认为教皇从根本上 误解意识是什么?
210Um, I think he thinks we were made by God.我觉得他觉得我们是上帝造的
211He is, after all, the pope.毕竟他是教皇
212And if he didn't think that, it'd have problems.如果他不这么认为,那会有问题。
213I don't believe we were made by God.我不相信我们是上帝造的
214I believe we were made by evolution.我相信我们是由进化而成的
215So of course the pope doesn't think these things could have consciousness教皇当然不会认为这些东西有意识
216or souls or whatever.或灵魂或什么的。
217I should say I have a lot of admiration for this pope.可以说我对这位教皇非常钦佩
218I think it's very important the work he's doing on how AI should be controlled.我认为他如何控制AI的工作非常重要.
219But I have a fundamental disagreement with him about the nature of beings.但我与他对于生命的本质有根本的分歧.
220I think with the product of evolution, he thinks we were made by God.我想用进化的产物 他以为我们是上帝造的
221It sounds like you're arguing that there is not something about听起来你好像在争论 没有什么
222being a living embodied human being that computation, however advanced, can't replace.人类是生命的象征 计算无论怎样先进 都无法取代
223That is what I'm arguing.这就是我的论点。
224I think consciousness is a property of a sophisticated kind of computation.我认为意识是一种复杂的计算方式
225And in fact, if you look at the scientists studying these AIs,事实上,如果你看看研究这些AI的科学家们
226when they're not thinking philosophically, when they're not thinking about the issue当他们不思考哲学, 当他们不思考问题
227of are they conscious, but just talking about what's going on,他们有意识,但只是谈论 发生了什么事,
228they say things in their papers like the AI was aware that it was being tested.他们在报纸上说的东西 像AI知道 它正在测试。
229Now, normally if you say that something's aware, that's attributing consciousness to it.现在,通常,如果你说某件事是意识到的, 那就是将意识归于它。
230So in everyday life, these AI scientists are attributing consciousness to these AIs.因此在日常生活中,这些AI科学家将意识归于这些AI.
231It's just when you start talking philosophy,只是当你开始谈论哲学,
232they get worried because they don't know what consciousness is.他们担心是因为不知道什么是意识
233If you believe that it's possible to perhaps instill within an AI a maternal instinct,如果你认为有可能在人工智能中注入母性的本能
234could AIs not also be trained on the values of the world's great religions?难道不能对AI进行世界各大宗教价值观的培训?
235Yes, they certainly could.是的,他们当然可以。
236So the way they're trained at present is they're trying to predict the next word in a document.因此,他们目前接受训练的方式是试图在文档中预测下一个词.
237And the documents they're given come from all sorts of different sources.而他们得到的文件来自各种不同的来源.
238They just scrape everything they can get their hands on for free on the web.他们只是刮了一切 他们可以得到的手 免费在网上。
239Now, if you were teaching your kid to read,如果你教孩子读书
240you wouldn't want to do that.你不会想那么做的
241I think the way they should be trained is they should only be trained我认为他们应该接受训练 只能接受训练
242on things that exhibit good behavior.对表现良好的东西。
243The kind of behavior we'd like these beings that we're going to share the planet with to exhibit.我们想让这些生物 共同展示地球的行为
244Once they've been trained on good behavior一旦他们接受了良好行为训练
245and they can recognize what good behavior is and what bad behavior is,他们可以识别什么是好行为 什么是坏行为,
246then it's fine to show them examples of bad behavior让他们看看不良行为的例子就好
247and tell them it's bad behavior.告诉他们这是坏行为
248They will figure out for themselves it's bad behavior.他们会发现这是坏行为
249You're suggesting we should be training them the way we ideally would train our children?你是说我们应该以我们最理想的方式训练他们?
250Exactly.没错
251So these AIs are not lines of code that we write.所以这些AI不是我们写的代码的线条.
252They're not software like normal computer software.它们不像普通的计算机软件那样是软件.
253They're things that learn from data.它们是学习数据的东西。
254And we would like them to learn to be good AIs in our definition of good.我们希望他们能学会成为好AI 在我们对好的定义。
255The control we have over them is the data we train them on我们控制他们,就是我们训练他们的数据
256and the rewards we give them, much like it is for children.和我们给他们的奖励, 就像它对于儿童一样。
257If we accept that AIs have the ability to have consciousness,如果我们承认AI有意识的能力
258what does that mean for our moral obligation towards an AI with consciousness?这对我们对有意识的AI的道德义务意味着什么?
259We're people, and what we care about is other people.我们是人,我们关心的是其他人。
260I don't see any reason why we shouldn't develop advanced intelligences我看不出为什么我们不发展高级情报
261that are smarter than us and that are conscious but have less moral rights than us.他们比我们聪明,他们有意识,但道德权利却比我们少。
262I would be perfectly happy with developing these things我很乐意开发这些东西
263so that they care about us and their view of the world is that we're what's really important所以他们关心我们和 他们的世界观 是,我们才是真正重要的
264and they're not so important.他们并不重要
265So I don't see why we should give them moral rights.所以我不明白我们为什么要给他们道德权利。
266One could imagine that we are engineering a form of digital slavery.人们可以想象,我们正在设计一种数字奴役形式。
267You could imagine that, yes.你能想象的到,是的。
268So we have to make a decision about this.因此,我们必须对此作出决定。
269And I think the decision we have to make if we produce superintelligence is,我认为如果我们生产超级智能,我们必须做的决定是,
270do we want to just be a passing phase in the evolution of intelligence我们是否只想成为 智力进化过程中的传承阶段
271that will eventually disappear, maybe quite soon?这最终会消失, 也许很快?
272Or do we want to try and stay in control, in which case these superintelligences或者我们想保持控制, 在这种情况下,这些超级情报
273that are smarter than us have to be made so what they really care about is us?所以他们真正关心的是我们?
274I'd like to turn to the news.转正报.
275AI巨头NVIDIA亿万富翁CEOJensen Wong辞职.
276Does leaving technology in the hands of private citizens技术是否由公民个人掌握
277that concerns the national security of nations,事关各国的国家安全,
278is that a risk we should be taking?我们应该冒这个险吗?
279Yes, it's obviously a risk.是的,这显然是一种风险。
280I need to say a few things about Jensen Wong.我要谈谈黄詹森的事
281Please do.请吧
282I actually like him.其实我很喜欢他
283I've interacted with him a bit.我和他有过互动
284I think he's a decent guy.我觉得他是个好人
285I think he made a mistake.我认为他犯了一个错误。
286I think it was slightly stupid to think he wouldn't be caught以为他不会被抓到,这有点傻
287because obviously the security services would be looking at somebody like him.因为很明显安全部门会 看着他这样的人
288I don't put him in the same category as someone like Musk or Zuckerberg,我不会把他和穆斯克或扎克伯格一样 归入同一类别
289who I think are very self-interested and willing to do more or less anything to make more money.我认为他很自私 愿意做多少事情来赚钱
290But obviously having the future of humanity controlled by just a few men is a bad idea.但显然只有几个人控制人类的未来是个坏主意。
291The reason I ask you about Jensen Wong is there is also this question of a race between我问你关于Jensen Wong的原因 也是关于比赛的问题
292the Chinese and the Western frontier labs and whether humanity will be in better shape中国和西方的前沿实验室 人类是否将处于更好的状态
293or more secure if a representative democracy reaches super intelligence or reaches,如果一个代议制民主国家达到超级智能或达到更安全的程度,
294continues to lead in the AI race.AI赛事继续领先.
295It's not unlike during the nuclear arms race.这与核军备竞赛期间的情况并无不同.
296It seems as though there is a similar dilemma we face with AI in this national security component似乎在国家安全方面 AI也有类似的两难处境
297of who ends up leading with this new advanced technology.最终领先于这种新的先进技术
298Okay, so when it comes to super intelligence,好吧,所以当涉及到超级智能,
299I think actually the interests of China and the U.S. are aligned and they will collaborate.我认为实际上中国和美国的利益是一致的,它们将会合作。
300It does not want super intelligence taking over.它不希望超级情报接管。
301Last summer I went to Shanghai and talked to a member of the Politburo.去年夏天我去了上海 和一位政治人物谈过
302I believe the Politburo understands this stuff much better than the U.S. government.我相信政治局比美国政府更了解这些
303They are very concerned about these risks and they will collaborate with the U.S. on trying他们非常关心这些风险 他们会与美国合作尝试
304to prevent super intelligence taking over.防止超级情报接管
305So I think what should be happening at present, it won't happen while we still have a,所以我认为现在应该发生什么 我们还有...
306how can I say this, maybe a corrupt senile felon in the White House.我怎么能这么说呢 也许白宫里有个腐败的老年重罪犯
307So it's not going to happen for two years, but we will eventually get a sensible U.S.所以两年内不会发生, 但我们最终会有一个明智的美国。
308that I think will collaborate with the Chinese on trying to avert the risk of super intelligence我想与中国人合作 避免超级智能的风险
309taking over, much as the United States collaborated with the Soviet Union in the 1950s on trying to1950年代美国与苏联合作试图
310avert the dangers created by H-bombs.避免H炸弹造成的危险
311One of the AI executives who you seem to trust the most is Dario Amode of Anthropic.你最信任的AI高管之一 是雅典的达里奥·阿莫德
312You gave him an A for his recent essay on why we should be pacing the AI frontier.你给了他一个A,因为他最近的文章 关于为什么我们应该 速度AI边境。
313Anthropic took steps this week towards one of the biggest本周,Anthropic采取了步骤 走向一个最大的
314IPOs in history with a valuation of $1.4 trillion.历史上的IPO估值为1.4万亿美元.
315Can even the most safety conscious AI companies put humanity's interest ahead of its shareholders即使最安全的AI公司 也能把人类的利益放在股东的首位
316when trillions of dollars are at stake?当万亿元处于危险之中时?
317So Dario is in a very difficult position.因此达里奥的处境非常困难.
318One thing I should say is don't look at what people say, look at what people do.我该说一件事,就是不要看别人说什么,看别人做什么.
319And so Claude was used by the U.S. military.于是克劳德被美国军方使用.
320The U.S. military wanted to be allowed to use it to spy on U.S. citizens.美国军方希望被允许利用来监视美国公民.
321And Dario said no and took a big hit for that.达里奥拒绝了,因此大受欢迎
322When Dario said no, Sam Altman said, we agree with him.当达里奥拒绝时 Sam Altman说 我们同意他的意见
323And then about a day after the U.S. military had fired Claude,在美军解雇克劳德的一天后
324Sam Altman说 好吧 打开AI 我来
325So Dario was willing to take a hit in order to do what he thought was right.于是达里奥为了做他认为对的事,愿意接受一击.
326That's much more convincing than anything they say.这比他们说的更有说服力
327How then, you know, if a company is going to market with an IPO valuation of $1.4 trillion如果一个公司要上市 IPO估值为1.4万亿美元
328and it says in its filing papers there is a 10% chance that civilization could be它说,在它的档案文件中 有10%的可能性 文明可能是
329at risk because of the technology that is going to market, help me understand why and how帮助我理解为什么和如何
330the marketplace is willing to accept this kind of risk.市场愿意接受这种风险。
331I mean, it would have been unheard of in the nuclear age when the Manhattan曼哈顿的核子时代 本来是没听说过的
332Project was being developed that something as dire as nuclear weapons正在开发一个像核武器一样可怕的项目
333would have been treated as as incautiously as the market seems to be treating this real risk.市场似乎在处理这种实际风险时会视之为不合理。
334From Dario's point of view, he has to make a decision about whether to say从达里奥的观点来看,他必须做出是否说
335this stuff shouldn't be developed, I'm out.这些东西不应该开发,我退出。
336In which case it's going to be developed by people who are less cautious than him.这样的话,就会由比他更谨慎的人来发展.
337That puts him in a very difficult situation, right?这让他陷入非常困难的境地,对吧?
338He has to try and stay up with the frontier so that if it's developed by anybody,他必须努力守住边境 如果它被任何人开发出来
339it's developed by Anthropic who are going to be more cautious than the other companies.它由Anthropic公司开发, 他们会比其他公司更谨慎。
340That was the original motivation of open AI.这是公开AI的最初动机.
341They said if anybody's going to develop super intelligence,他们说如果有人要发展超级智能
342it should be developed by somebody who really cares about safety.它应该由真正关心安全的人来开发。
343And open AI was set up to really care about safety.开放AI是为了真正关心安全
344Since then, it's started to care much less about safety.从那时起,它开始更不关心安全.
345So I think from Dario's point of view, the reason he's in a bind is所以我认为从达里奥的观点来看 他陷入困境的原因是
346he's very worried about the safety,他非常担心安全
347but he thinks it's better if Anthropic develops it than other companies.但他认为如果Anthropic开发它比其他公司更好.
348And I know quite a few people who work for Anthropic precisely because我认识很多为Anthropic工作的人
349they think it's going to be the safest company for developing AI.他们认为这将是开发AI最安全的公司.
350This summer, your fellow AI pioneer and Nobel laureate, Demis Hassibis,今年夏天 你们的AI先驱和诺贝尔奖得主 戴米斯·哈西比斯
351outlined a proposal for federally overseen self-regulatory organization概述关于联邦监督自律组织的建议
352with a board that includes a set of independent technical experts.设有一个包括一组独立技术专家的委员会。
353The Wall Street Journal reported that some in the White House were receptive to the idea,"华尔街日报"报道,白宫的一些人接受这个想法,
354but AI executives like Mark Zuckerberg, Elon Musk, and even Jensen Huang包括马克·扎克伯格、伊隆·穆斯克 甚至黄詹森
355encouraged President Trump to reject the plan.鼓励特朗普总统拒绝该计划.
356Is that a workable solution still to your way of thinking?你的思想方式还有这个可行办法吗?
357I think it's a first step.我认为这是第一步。
358It's not the whole solution,这不是整个解决方案,
359but I think it's crazy that at present there's almost no regulation on AI.但我认为这太疯狂了 目前几乎没有关于AI的规定。
360The regulation we have on pharmaceutical drugs, for example,比如,我们有关药品的条例,
361is that if you want to release a new pharmaceutical drug,如果你想释放一种新药
362you have to convince the FDA that it's safe,你必须让FDA相信这是安全的,
363and you have to do the work to convince the FDA,你必须努力说服FDA
364and the FDA can consult external experts.林业发展局可以咨询外部专家。
365We need something like that for AI.我们需要这样的东西对AI。
366That's the very minimum we need,这是我们最需要的
367and it's crazy we don't have it.我们没有它,这是疯了。
368This program is the revival of William F. Buckley Jr.'s Firing Line,这个计划是恢复小威廉·F·巴克利的射击线
369which aired for 33 years on PBS.在PBS播出了33年
370In 1993, William F. Buckley Jr. hosted a discussion1993年,小威廉·F·巴克利主持讨论.
371about the proliferation of nuclear weapons and the threat of nuclear destruction.核武器的扩散和核毁灭的威胁。
372Here is a physicist, Russell Seitz,这是物理学家罗素·塞兹
373talking about the challenge of containing nuclear technology.谈论遏制核技术的挑战。
374As high technology ramifies appropriate technologies由于高技术是适当的技术
375that can be subverted to the ends of proliferation multiply,被破坏到扩散的目的,
376until at the margin, we may be forced to rely more on our powers of diplomacy在边上,我们可能不得不更多地依赖我们的外交力量
377than on our capacity to suppress the progress of technology globally,而不是我们抑制全球技术进步的能力,
378if we are to keep under control the prospect of essentially unlimited proliferation of nuclear weapons.如果我们要控制核武器基本上无限扩散的前景。
379Okay, so there are some big differences between AI and nuclear weapons.好吧,所以在AI和核武器之间有一些很大的区别。
380The first big difference is AI can do tremendous good.第一个大区别是AI可以做巨大的好事.
381We haven't talked about that,我们还没谈这个
382but it'll do tremendous good in healthcare and education,但它会在保健和教育方面产生巨大效果,
383and in making almost every industry more productive.让几乎所有的行业都更加富有生产力。
384So it's going to be developed for that reason.因此,它会为此发展起来.
385We can't stop it, I don't think.我们不能阻止它,我不认为。
386But the second thing is nuclear weapons aren't intelligent.但第二件事是核武器不聪明。
387AI is intelligent, and if we can educate it and train itAI很聪明,如果我们能教育它,训练它
388so that it really cares about us, then proliferation is not nearly so worrying.这样它才真正关心我们,那么扩散就不那么令人担忧。
389You argue the technology that you have spent your life developing你争辩你一生的科技发展
390might have been a mistake.也许是个错误
391And you address this head on in your forthcoming book,在你即将出版的书中 写着这个头
392writing, quote, responsibility is not the same thing as regret.写作、引用、责任与后悔不同。
393Explain the difference.解释差别.
394Okay, I want to distinguish actually two kinds of regret.好吧,我想区分两种悔恨
395Sort of sometimes you're mean to a friend of yours,有时候你对你的朋友很刻薄
396and you regret it.你却后悔了
397That's what I call guilty regret.这就是我所谓的有罪遗憾。
398When you did something and at the time you knew it was wrong,当你做了一些事情 当你知道这是错的,
399but you did it anyway.但你还是做到了
400Then there's things that you did where at the time然后,有些事情 你做了在当时
401you couldn't have foreseen the consequences,你无法预见后果
402but they turned out to have potentially very bad consequences.但结果却产生了非常糟糕的后果
403I feel that that's what I did.我觉得这就是我做了什么。
404And if I was put in the same situation again如果我再次陷入同样的境地
405with the knowledge I had at the time, I would have behaved the same way.以我当时的知识,我也会采取同样的做法。
406Dr. Hinton, thank you so much for joining me on Firing Line.欣顿博士,非常感谢你 加入我Firing线。
407Thank you for inviting me.谢谢你邀请我
408Firing Line with Margaret Hoover is made possible in part by与玛格丽特·胡佛的交火线部分通过
409Robert Granieri, Vanessa and Henry Cornell,罗伯特·格拉尼埃里 瓦内萨和亨利·康奈尔
410the Margaret and Daniel Loeb Foundation,玛格丽特和丹尼尔·洛布基金会,
411the Beth and Ravenel Curry Foundation,贝丝和雷文埃尔·库里基金会,
412Cliff and Laurel Astes,克里夫和劳蕾尔·阿斯特斯
413and by the following.以及以下内容。
414你看着PBS。
该视频共有字幕 414 条。解锁更多字幕为会员功能,请移动到 价格