TED2015
Nick Bostrom: What happens when our computers get smarter than we are?
尼克·博斯特罗姆: 当电脑比我们还聪明时会发生什么？
Filmed:
Readability: 4.4
4,632,705 views
人工智能技术在这个世纪中突飞猛进，研究表明，人工智能电脑可以和人类一样聪明。而在那之后，尼克·博斯特罗姆认为，它们将会超越我们：“机器智能是人类的最终发明。” 作为一个科技学家和哲学家，博斯特罗姆要我们反思我们正在建造的世界，这世界由具有思考能力的机器所主宰。这些聪明的机器会帮助我们维持人性和价值观？还是会创造出它们自己的价值观？
Nick Bostrom - Philosopher
Nick Bostrom asks big questions: What should we do, as individuals and as a species, to optimize our long-term prospects? Will humanity’s technological advancements ultimately destroy us? Full bio
Nick Bostrom asks big questions: What should we do, as individuals and as a species, to optimize our long-term prospects? Will humanity’s technological advancements ultimately destroy us? Full bio
Double-click the English transcript below to play the video.
00:12
I work with a bunch of mathematicians,
philosophers and computer scientists,
philosophers and computer scientists,
0
570
4207
我和一些数学家、
哲学家和电脑学家一起工作，
哲学家和电脑学家一起工作，
00:16
and we sit around and think about
the future of machine intelligence,
the future of machine intelligence,
1
4777
5209
我们会坐在一起思考未来的机械智能，
00:21
among other things.
2
9986
2044
和其他的一些事情。
00:24
Some people think that some of these
things are sort of science fiction-y,
things are sort of science fiction-y,
3
12030
4725
有的人认为这类事情只是科幻，
00:28
far out there, crazy.
4
16755
3101
不切实际，很疯狂。
00:31
But I like to say,
5
19856
1470
但是我想说，
00:33
okay, let's look at the modern
human condition.
human condition.
6
21326
3604
好吧，那我们来看看人类现状吧。
00:36
(Laughter)
7
24930
1692
（笑）
00:38
This is the normal way for things to be.
8
26622
2402
这是世间一种常态。
00:41
But if we think about it,
9
29024
2285
但是如果我们去思考，
00:43
we are actually recently arrived
guests on this planet,
guests on this planet,
10
31309
3293
我们人类，其实相当晚才
00:46
the human species.
11
34602
2082
出现在这个星球上。
00:48
Think about if Earth
was created one year ago,
was created one year ago,
12
36684
4746
想一想，如果地球是一年前才被创造的，
00:53
the human species, then,
would be 10 minutes old.
would be 10 minutes old.
13
41430
3548
人类，那么，10分钟前才出现。
00:56
The industrial era started
two seconds ago.
two seconds ago.
14
44978
3168
然后工业时代两秒钟前刚刚开始。
01:01
Another way to look at this is to think of
world GDP over the last 10,000 years,
world GDP over the last 10,000 years,
15
49276
5225
另一种看待这件事的方式是去
想一下在过去一万年间的世界 GDP 状况。
想一下在过去一万年间的世界 GDP 状况。
01:06
I've actually taken the trouble
to plot this for you in a graph.
to plot this for you in a graph.
16
54501
3029
我其实真的试着去做了一个统计图。
01:09
It looks like this.
17
57530
1774
就是这样。
01:11
(Laughter)
18
59304
1363
（笑）
01:12
It's a curious shape
for a normal condition.
for a normal condition.
19
60667
2151
这是个令人好奇的形状，正常情况下。
01:14
I sure wouldn't want to sit on it.
20
62818
1698
我确定我不想坐在上面。
01:16
(Laughter)
21
64516
2551
（笑）
01:19
Let's ask ourselves, what is the cause
of this current anomaly?
of this current anomaly?
22
67067
4774
让我们扪心自问，到底是什么造成了
如此不寻常的现状？
如此不寻常的现状？
01:23
Some people would say it's technology.
23
71841
2552
一些人会说，因为科技。
01:26
Now it's true, technology has accumulated
through human history,
through human history,
24
74393
4668
对于现在来说是对的，科技是人类历史
不断积累下来的果实。
不断积累下来的果实。
01:31
and right now, technology
advances extremely rapidly --
advances extremely rapidly --
25
79061
4652
现在，科技发展十分迅速：
01:35
that is the proximate cause,
26
83713
1565
这是个直接原因，
01:37
that's why we are currently
so very productive.
so very productive.
27
85278
2565
这就是为什么我们现在生产效率如此高。
01:40
But I like to think back further
to the ultimate cause.
to the ultimate cause.
28
88473
3661
但是我想探究更远的在未来的终极原因。
01:45
Look at these two highly
distinguished gentlemen:
distinguished gentlemen:
29
93114
3766
看这两个非常不同的男士：
01:48
We have Kanzi --
30
96880
1600
这是 Kanzi，
01:50
he's mastered 200 lexical
tokens, an incredible feat.
tokens, an incredible feat.
31
98480
4643
他已经掌握了 200 个词法标记，
一个难以置信的成就。
一个难以置信的成就。
01:55
And Ed Witten unleashed the second
superstring revolution.
superstring revolution.
32
103123
3694
Ed Witten 开创了第二个令人惊人的创新。
01:58
If we look under the hood,
this is what we find:
this is what we find:
33
106817
2324
如果我们去看这些事物的本质，
这是我们的发现：
这是我们的发现：
02:01
basically the same thing.
34
109141
1570
全都是一样的。
02:02
One is a little larger,
35
110711
1813
一个稍微大了一点，
02:04
it maybe also has a few tricks
in the exact way it's wired.
in the exact way it's wired.
36
112524
2758
也许它有一些特殊的技巧。
02:07
These invisible differences cannot
be too complicated, however,
be too complicated, however,
37
115282
3812
但是，这些隐形的不同并没有很错综复杂，
02:11
because there have only
been 250,000 generations
been 250,000 generations
38
119094
4285
因为在我们和我们的祖先之间
02:15
since our last common ancestor.
39
123379
1732
只有 25 万代人。
02:17
We know that complicated mechanisms
take a long time to evolve.
take a long time to evolve.
40
125111
3849
我们知道复杂的机制
需要很长的时间来进化得到。
需要很长的时间来进化得到。
02:22
So a bunch of relatively minor changes
41
130000
2499
所以，一些相对小的变化，
02:24
take us from Kanzi to Witten,
42
132499
3067
让我们从 Kanzi 变成了 Witten，
02:27
from broken-off tree branches
to intercontinental ballistic missiles.
to intercontinental ballistic missiles.
43
135566
4543
从捡起掉下的树枝作为武器，
到发射洲际导弹。
到发射洲际导弹。
02:32
So this then seems pretty obvious
that everything we've achieved,
that everything we've achieved,
44
140839
3935
因此，至今我们所办到的所有事情，
02:36
and everything we care about,
45
144774
1378
以及我们所关心的事情，
02:38
depends crucially on some relatively minor
changes that made the human mind.
changes that made the human mind.
46
146152
5228
都取决于人大脑中细小的变化。
02:44
And the corollary, of course,
is that any further changes
is that any further changes
47
152650
3662
因此得出的结论是：在未来，
02:48
that could significantly change
the substrate of thinking
the substrate of thinking
48
156312
3477
任何显著的思考基体的变化，
02:51
could have potentially
enormous consequences.
enormous consequences.
49
159789
3202
都能带来巨大的后果。
02:56
Some of my colleagues
think we're on the verge
think we're on the verge
50
164321
2905
我的一些同事觉得我们即将会发明，
02:59
of something that could cause
a profound change in that substrate,
a profound change in that substrate,
51
167226
3908
足以深深地改变人类思考模式的科技。
03:03
and that is machine superintelligence.
52
171134
3213
就是超级机能智慧。
03:06
Artificial intelligence used to be
about putting commands in a box.
about putting commands in a box.
53
174347
4739
以前的人工智慧
是把指令输入到一个箱子里。
是把指令输入到一个箱子里。
03:11
You would have human programmers
54
179086
1665
你需要人类程序员，
03:12
that would painstakingly
handcraft knowledge items.
handcraft knowledge items.
55
180751
3135
来努力把知识转变成程序。
03:15
You build up these expert systems,
56
183886
2086
你会建立起一些专业系统，
03:17
and they were kind of useful
for some purposes,
for some purposes,
57
185972
2324
它们有时候会有帮助，
03:20
but they were very brittle,
you couldn't scale them.
you couldn't scale them.
58
188296
2681
但是它们很生硬，你不能延展它们的功能。
03:22
Basically, you got out only
what you put in.
what you put in.
59
190977
3433
基本上你只能得到你放进去的东西。
03:26
But since then,
60
194410
997
但是自从那时候开始，
03:27
a paradigm shift has taken place
in the field of artificial intelligence.
in the field of artificial intelligence.
61
195407
3467
人工智能的领域发生了巨大的改变。
03:30
Today, the action is really
around machine learning.
around machine learning.
62
198874
2770
现在主要的研究方向是机器的学习。
03:34
So rather than handcrafting knowledge
representations and features,
representations and features,
63
202394
5387
所以，预期设计出知识的再现，
03:40
we create algorithms that learn,
often from raw perceptual data.
often from raw perceptual data.
64
208511
5554
我们写出具有从原始感官数据学习的程序，
03:46
Basically the same thing
that the human infant does.
that the human infant does.
65
214065
4998
像婴儿一样。
03:51
The result is A.I. that is not
limited to one domain --
limited to one domain --
66
219063
4207
结果就不会局限于某个领域的人工智能：
03:55
the same system can learn to translate
between any pairs of languages,
between any pairs of languages,
67
223270
4631
同一个系统可以学习两种语言之间的翻译
03:59
or learn to play any computer game
on the Atari console.
on the Atari console.
68
227901
5437
或者学着玩 Atari 的游戏。
04:05
Now of course,
69
233338
1779
当然，现在，
04:07
A.I. is still nowhere near having
the same powerful, cross-domain
the same powerful, cross-domain
70
235117
3999
人工智能还未能达到向人类一样，
04:11
ability to learn and plan
as a human being has.
as a human being has.
71
239116
3219
具有强大的跨领域学习能力。
04:14
The cortex still has some
algorithmic tricks
algorithmic tricks
72
242335
2126
人类大脑还具有一些运算技巧，
04:16
that we don't yet know
how to match in machines.
how to match in machines.
73
244461
2355
可是我们不知道如何
将这些技巧用于机器。
将这些技巧用于机器。
04:19
So the question is,
74
247886
1899
所以我们现在需要问：
04:21
how far are we from being able
to match those tricks?
to match those tricks?
75
249785
3500
我们还要多久才可以
让机器复制这种能力？
让机器复制这种能力？
04:26
A couple of years ago,
76
254245
1083
几年前，
04:27
we did a survey of some of the world's
leading A.I. experts,
leading A.I. experts,
77
255328
2888
我们对世界顶尖的人工智能专家
做了一次问卷调查，
做了一次问卷调查，
04:30
to see what they think,
and one of the questions we asked was,
and one of the questions we asked was,
78
258216
3224
来收集他们的想法，其中一道题目是：
04:33
"By which year do you think
there is a 50 percent probability
there is a 50 percent probability
79
261440
3353
“到哪一年你觉得人类会有 50% 的可能性
04:36
that we will have achieved
human-level machine intelligence?"
human-level machine intelligence?"
80
264793
3482
创造达到人类水平的人工智能？”
04:40
We defined human-level here
as the ability to perform
as the ability to perform
81
268785
4183
我们把这样的人工智能定义为
04:44
almost any job at least as well
as an adult human,
as an adult human,
82
272968
2871
有能力将任何任务
完成得至少和一名成年人一样好。
完成得至少和一名成年人一样好。
04:47
so real human-level, not just
within some limited domain.
within some limited domain.
83
275839
4005
所以是真正的人类级别，
而不是仅限于一些领域。
而不是仅限于一些领域。
04:51
And the median answer was 2040 or 2050,
84
279844
3650
而答案的中位数是 2040 到 2050 年，
04:55
depending on precisely which
group of experts we asked.
group of experts we asked.
85
283494
2806
取决于这些专家的群体。
04:58
Now, it could happen much,
much later, or sooner,
much later, or sooner,
86
286300
4039
当然这个有可能要过很久才能实现，
也有可能提前实现。
也有可能提前实现。
05:02
the truth is nobody really knows.
87
290339
1940
没有人知道确切的时间。
05:05
What we do know is that the ultimate
limit to information processing
limit to information processing
88
293259
4412
我们知道的事，
处理信息的能力的最终点，
处理信息的能力的最终点，
05:09
in a machine substrate lies far outside
the limits in biological tissue.
the limits in biological tissue.
89
297671
4871
比任何生物组织要大很多。
05:15
This comes down to physics.
90
303241
2378
这取决与物理原理。
05:17
A biological neuron fires, maybe,
at 200 hertz, 200 times a second.
at 200 hertz, 200 times a second.
91
305619
4718
一个生物神经元所发出的脉冲频率
大约位于 200 赫兹，每秒 200 次。
大约位于 200 赫兹，每秒 200 次。
05:22
But even a present-day transistor
operates at the Gigahertz.
operates at the Gigahertz.
92
310337
3594
但是就算是现在的电晶体
都以千兆赫的频率运行。
都以千兆赫的频率运行。
05:25
Neurons propagate slowly in axons,
100 meters per second, tops.
100 meters per second, tops.
93
313931
5297
神经元在轴突中传输的速度较慢，
最多 100 米每秒。
最多 100 米每秒。
05:31
But in computers, signals can travel
at the speed of light.
at the speed of light.
94
319228
3111
但在电脑里，信号是以光速传播的。
05:35
There are also size limitations,
95
323079
1869
另外还有尺寸的限制，
05:36
like a human brain has
to fit inside a cranium,
to fit inside a cranium,
96
324948
3027
就像人类的大脑只能有颅骨那么大，
05:39
but a computer can be the size
of a warehouse or larger.
of a warehouse or larger.
97
327975
4761
但是一个电脑可以和仓库一样大，甚至更大。
05:44
So the potential for superintelligence
lies dormant in matter,
lies dormant in matter,
98
332736
5599
因此超级智慧的潜能正潜伏在物质之中，
05:50
much like the power of the atom
lay dormant throughout human history,
lay dormant throughout human history,
99
338335
5712
就像原子能潜伏在人类历史中一样，
05:56
patiently waiting there until 1945.
100
344047
4405
直到 1945。
06:00
In this century,
101
348452
1248
在这个世纪里，
06:01
scientists may learn to awaken
the power of artificial intelligence.
the power of artificial intelligence.
102
349700
4118
科学家可能能将人工智慧的力量唤醒。
06:05
And I think we might then see
an intelligence explosion.
an intelligence explosion.
103
353818
4008
那时候我觉得我们会看到智慧大爆发。
06:10
Now most people, when they think
about what is smart and what is dumb,
about what is smart and what is dumb,
104
358406
3957
大部分的人，当他们想
什么是聪明什么是笨的时候，
什么是聪明什么是笨的时候，
06:14
I think have in mind a picture
roughly like this.
roughly like this.
105
362363
3023
他们脑子里的画面是这样的：
06:17
So at one end we have the village idiot,
106
365386
2598
一边是村子里的傻子，
06:19
and then far over at the other side
107
367984
2483
一边是
06:22
we have Ed Witten, or Albert Einstein,
or whoever your favorite guru is.
or whoever your favorite guru is.
108
370467
4756
Ed Witten 或 Albert Einstein，
或者其他大师。
或者其他大师。
06:27
But I think that from the point of view
of artificial intelligence,
of artificial intelligence,
109
375223
3834
但是我觉得从人工智能的观点来看，
06:31
the true picture is actually
probably more like this:
probably more like this:
110
379057
3681
真正的画面也许是这样：
06:35
AI starts out at this point here,
at zero intelligence,
at zero intelligence,
111
383258
3378
人工智能从这一点开始，零智慧。
06:38
and then, after many, many
years of really hard work,
years of really hard work,
112
386636
3011
然后，在许多许多辛劳工作后，
06:41
maybe eventually we get to
mouse-level artificial intelligence,
mouse-level artificial intelligence,
113
389647
3844
也许最终我们能达到老鼠级别的智慧，
06:45
something that can navigate
cluttered environments
cluttered environments
114
393491
2430
能在混乱中找到开出一条道路，
06:47
as well as a mouse can.
115
395921
1987
像一只老鼠一样。
06:49
And then, after many, many more years
of really hard work, lots of investment,
of really hard work, lots of investment,
116
397908
4313
之后，在更多更多年的辛苦研究
和投资之后，
和投资之后，
06:54
maybe eventually we get to
chimpanzee-level artificial intelligence.
chimpanzee-level artificial intelligence.
117
402221
4639
也许最终我们能到达黑猩猩级人工智能。
06:58
And then, after even more years
of really, really hard work,
of really, really hard work,
118
406860
3210
在后来，更多年的研究之后，
07:02
we get to village idiot
artificial intelligence.
artificial intelligence.
119
410070
2913
我们能够达到村里的傻子级别的人工智能。
07:04
And a few moments later,
we are beyond Ed Witten.
we are beyond Ed Witten.
120
412983
3272
在一段时间之后，
我们能超越 Ed Witten。
我们能超越 Ed Witten。
07:08
The train doesn't stop
at Humanville Station.
at Humanville Station.
121
416255
2970
这列火车不会在“人类站”就停下。
07:11
It's likely, rather, to swoosh right by.
122
419225
3022
它比较可能会呼啸而过。
07:14
Now this has profound implications,
123
422247
1984
现在这个有深远的寓意，
07:16
particularly when it comes
to questions of power.
to questions of power.
124
424231
3862
尤其是当我们谈到力量权利的时候。
07:20
For example, chimpanzees are strong --
125
428093
1899
比如，黑猩猩很强壮：
07:21
pound for pound, a chimpanzee is about
twice as strong as a fit human male.
twice as strong as a fit human male.
126
429992
5222
同等的体重，一个黑猩猩是
两个健康男性那么强壮。
两个健康男性那么强壮。
07:27
And yet, the fate of Kanzi
and his pals depends a lot more
and his pals depends a lot more
127
435214
4614
然而，Kanzi 和他的朋友们的命运
更多取决于
更多取决于
07:31
on what we humans do than on
what the chimpanzees do themselves.
what the chimpanzees do themselves.
128
439828
4140
我们人类能做到什么，
而不是猩猩能做到什么。
而不是猩猩能做到什么。
07:37
Once there is superintelligence,
129
445228
2314
当超级智慧出现的时候，
07:39
the fate of humanity may depend
on what the superintelligence does.
on what the superintelligence does.
130
447542
3839
人类的命运也许会取决于
那个超级智慧体要做什么。
那个超级智慧体要做什么。
07:44
Think about it:
131
452451
1057
想一想：
07:45
Machine intelligence is the last invention
that humanity will ever need to make.
that humanity will ever need to make.
132
453508
5044
机器智慧是人类需要创造的最后一个东西。
07:50
Machines will then be better
at inventing than we are,
at inventing than we are,
133
458552
2973
机器在那之后会比我们更擅长创造，
07:53
and they'll be doing so
on digital timescales.
on digital timescales.
134
461525
2540
他们也会在数位时间里这样做。
07:56
What this means is basically
a telescoping of the future.
a telescoping of the future.
135
464065
4901
这个意味着一个被缩短的未来。
08:00
Think of all the crazy technologies
that you could have imagined
that you could have imagined
136
468966
3558
想一下你曾想象过的所有的疯狂的科技，
08:04
maybe humans could have developed
in the fullness of time:
in the fullness of time:
137
472524
2798
也许人类可以在适当的时候完成：
08:07
cures for aging, space colonization,
138
475322
3258
终结衰老、宇宙殖民、
08:10
self-replicating nanobots or uploading
of minds into computers,
of minds into computers,
139
478580
3731
自我复制的纳米机器人
和大脑到电脑的传输，
和大脑到电脑的传输，
08:14
all kinds of science fiction-y stuff
140
482311
2159
诸如此类的看似仅存在于科幻
08:16
that's nevertheless consistent
with the laws of physics.
with the laws of physics.
141
484470
2737
却有同时符合物理法则的元素。
08:19
All of this superintelligence could
develop, and possibly quite rapidly.
develop, and possibly quite rapidly.
142
487207
4212
超级智慧有办法开发出这些东西，也许更快。
08:24
Now, a superintelligence with such
technological maturity
technological maturity
143
492449
3558
现在，一个拥有如此成熟科技的超级智慧体
08:28
would be extremely powerful,
144
496007
2179
将会是非常强大，
08:30
and at least in some scenarios,
it would be able to get what it wants.
it would be able to get what it wants.
145
498186
4546
至少在一些情况下，
它能得到它想要的东西。
它能得到它想要的东西。
08:34
We would then have a future that would
be shaped by the preferences of this A.I.
be shaped by the preferences of this A.I.
146
502732
5661
我们的未来就将会被
这个超级智慧体的喜好所主宰。
这个超级智慧体的喜好所主宰。
08:41
Now a good question is,
what are those preferences?
what are those preferences?
147
509855
3749
现在的问题就是，
这些喜好是什么呢？
这些喜好是什么呢？
08:46
Here it gets trickier.
148
514244
1769
这很棘手。
08:48
To make any headway with this,
149
516013
1435
要在这个领域取得进步，
08:49
we must first of all
avoid anthropomorphizing.
avoid anthropomorphizing.
150
517448
3276
我们必须避免将机器智慧人格化。
08:53
And this is ironic because
every newspaper article
every newspaper article
151
521934
3301
这一点很讽刺，
因为每一个关于人工智能的未来
因为每一个关于人工智能的未来
08:57
about the future of A.I.
has a picture of this:
has a picture of this:
152
525235
3855
的新闻报道，都会有这个图片：
09:02
So I think what we need to do is
to conceive of the issue more abstractly,
to conceive of the issue more abstractly,
153
530280
4134
所以我觉得我们必须
用更抽象的方法看待这个问题，
用更抽象的方法看待这个问题，
09:06
not in terms of vivid Hollywood scenarios.
154
534414
2790
而不是在好莱坞电影的叙事之下。
09:09
We need to think of intelligence
as an optimization process,
as an optimization process,
155
537204
3617
我们需要把智慧看做是一个优化的过程，
09:12
a process that steers the future
into a particular set of configurations.
into a particular set of configurations.
156
540821
5649
一个能把未来引导至
一个特殊组合结构的过程。
一个特殊组合结构的过程。
09:18
A superintelligence is
a really strong optimization process.
a really strong optimization process.
157
546470
3511
一个超级智慧体是一个
非常强大的优化过程。
非常强大的优化过程。
09:21
It's extremely good at using
available means to achieve a state
available means to achieve a state
158
549981
4117
它将会擅长利用资源来
达到自己的目标。
09:26
in which its goal is realized.
159
554098
1909
09:28
This means that there is no necessary
conenction between
conenction between
160
556447
2672
这意味着有着高智慧和
09:31
being highly intelligent in this sense,
161
559119
2734
拥有一个对人类来说有用的目标之间
09:33
and having an objective that we humans
would find worthwhile or meaningful.
would find worthwhile or meaningful.
162
561853
4662
并没有必然的联系。
09:39
Suppose we give an A.I. the goal
to make humans smile.
to make humans smile.
163
567321
3794
假设我们给予人工智慧的目的是让人笑，
09:43
When the A.I. is weak, it performs useful
or amusing actions
or amusing actions
164
571115
2982
当人工智能弱的时候，
它能做出有用或好笑的表演，
它能做出有用或好笑的表演，
09:46
that cause its user to smile.
165
574097
2517
这样它的使用者就会笑了。
09:48
When the A.I. becomes superintelligent,
166
576614
2417
当人工智能变成超级智慧体的时候，
09:51
it realizes that there is a more
effective way to achieve this goal:
effective way to achieve this goal:
167
579031
3523
它会意识到有一个更有效的办法
能达到这个效果：
能达到这个效果：
09:54
take control of the world
168
582554
1922
控制世界，
09:56
and stick electrodes into the facial
muscles of humans
muscles of humans
169
584476
3162
在人类面部肌肉上插入电极
09:59
to cause constant, beaming grins.
170
587638
2941
来让人类不断地笑。
10:02
Another example,
171
590579
1035
另一个例子：
10:03
suppose we give A.I. the goal to solve
a difficult mathematical problem.
a difficult mathematical problem.
172
591614
3383
假设我们给予人工智能的目标
是解出很难的数学题，
是解出很难的数学题，
10:06
When the A.I. becomes superintelligent,
173
594997
1937
当人工智能变成超级智慧体的时候，
10:08
it realizes that the most effective way
to get the solution to this problem
to get the solution to this problem
174
596934
4171
它意识到有一个更有效的办法来解出问题，
10:13
is by transforming the planet
into a giant computer,
into a giant computer,
175
601105
2930
是把整个地球变成一个巨型电脑，
10:16
so as to increase its thinking capacity.
176
604035
2246
这样它的运算能力就变更强大了。
10:18
And notice that this gives the A.I.s
an instrumental reason
an instrumental reason
177
606281
2764
注意到这个是
给予人工智能一个模式型的理由
给予人工智能一个模式型的理由
10:21
to do things to us that we
might not approve of.
might not approve of.
178
609045
2516
来做我们也许并不认可的事情。
10:23
Human beings in this model are threats,
179
611561
1935
人类在这个模式中是威胁，
10:25
we could prevent the mathematical
problem from being solved.
problem from being solved.
180
613496
2921
我们可以人为地
让这个数学问题不能被解出。
让这个数学问题不能被解出。
10:29
Of course, perceivably things won't
go wrong in these particular ways;
go wrong in these particular ways;
181
617207
3494
当然了，我们预见
这种事情不会错到这样的地步，
这种事情不会错到这样的地步，
10:32
these are cartoon examples.
182
620701
1753
这些是夸张的例子。
10:34
But the general point here is important:
183
622454
1939
但是它们所代表的主旨很重要：
10:36
if you create a really powerful
optimization process
optimization process
184
624393
2873
如果你创造了一个非常强大的优化过程
10:39
to maximize for objective x,
185
627266
2234
来最大化目标 X，
10:41
you better make sure
that your definition of x
that your definition of x
186
629500
2276
你最好保证你的意义上的 X
10:43
incorporates everything you care about.
187
631776
2469
包括了所有你在乎的事情。
10:46
This is a lesson that's also taught
in many a myth.
in many a myth.
188
634835
4384
这是一个很多神话故事中都在传递的寓意。
10:51
King Midas wishes that everything
he touches be turned into gold.
he touches be turned into gold.
189
639219
5298
（希腊神话中）的 Midas 国王
希望他碰到的所有东西都能变成金子。
希望他碰到的所有东西都能变成金子。
10:56
He touches his daughter,
she turns into gold.
she turns into gold.
190
644517
2861
他碰到了他的女儿，她于是变成了金子。
10:59
He touches his food, it turns into gold.
191
647378
2553
他碰到了食物，于是食物变成了金子。
11:01
This could become practically relevant,
192
649931
2589
这个故事和我们的话题息息相关，
11:04
not just as a metaphor for greed,
193
652520
2070
并不只是因为它隐藏在对贪婪的暗喻，
11:06
but as an illustration of what happens
194
654590
1895
也是因为他指出了
11:08
if you create a powerful
optimization process
optimization process
195
656485
2837
如果你创造出来一个强大的优化过程
11:11
and give it misconceived
or poorly specified goals.
or poorly specified goals.
196
659322
4789
并且给他了一个错误的或者不精确的目标，
后果会是什么。
后果会是什么。
11:16
Now you might say, if a computer starts
sticking electrodes into people's faces,
sticking electrodes into people's faces,
197
664111
5189
现在也许你会说，
如果一个电脑开始在人类脸上插电极，
如果一个电脑开始在人类脸上插电极，
11:21
we'd just shut it off.
198
669300
2265
我们会关掉它。
11:24
A, this is not necessarily so easy to do
if we've grown dependent on the system --
if we've grown dependent on the system --
199
672555
5340
第一，这不是一件容易事，
如果我们变得非常依赖这个系统：
如果我们变得非常依赖这个系统：
11:29
like, where is the off switch
to the Internet?
to the Internet?
200
677895
2732
比如，你知道互联网的开关在哪吗？
11:32
B, why haven't the chimpanzees
flicked the off switch to humanity,
flicked the off switch to humanity,
201
680627
5120
第二，为什么当初黑猩猩
没有关掉人类的开关，
没有关掉人类的开关，
11:37
or the Neanderthals?
202
685747
1551
或者尼安德特人的开关？
11:39
They certainly had reasons.
203
687298
2666
他们肯定有理由。
11:41
We have an off switch,
for example, right here.
for example, right here.
204
689964
2795
我们有一个开关，比如，这里。
11:44
(Choking)
205
692759
1554
（窒息声）
11:46
The reason is that we are
an intelligent adversary;
an intelligent adversary;
206
694313
2925
之所以我们是聪明的敌人，
11:49
we can anticipate threats
and plan around them.
and plan around them.
207
697238
2728
因为我们可以预见到威胁并且尝试避免它。
11:51
But so could a superintelligent agent,
208
699966
2504
但是一个超级智慧体也可以，
11:54
and it would be much better
at that than we are.
at that than we are.
209
702470
3254
而且会做得更好。
11:57
The point is, we should not be confident
that we have this under control here.
that we have this under control here.
210
705724
7187
我们不应该很自信地
表示我们能控制所有事情。
表示我们能控制所有事情。
12:04
And we could try to make our job
a little bit easier by, say,
a little bit easier by, say,
211
712911
3447
为了把我们的工作变得更简单一点，
我们应该试着，比如，
我们应该试着，比如，
12:08
putting the A.I. in a box,
212
716358
1590
把人工智能放进一个小盒子，
12:09
like a secure software environment,
213
717948
1796
想一个保险的软件环境，
12:11
a virtual reality simulation
from which it cannot escape.
from which it cannot escape.
214
719744
3022
一个它无法逃脱的虚拟现实模拟器。
12:14
But how confident can we be that
the A.I. couldn't find a bug.
the A.I. couldn't find a bug.
215
722766
4146
但是我们有信心它不可能能发现一个漏洞，
能让它逃出的漏洞吗？
能让它逃出的漏洞吗？
12:18
Given that merely human hackers
find bugs all the time,
find bugs all the time,
216
726912
3169
连人类黑客每时每刻都能发现网络漏洞，
12:22
I'd say, probably not very confident.
217
730081
3036
我会说，也许不是很有信心。
12:26
So we disconnect the ethernet cable
to create an air gap,
to create an air gap,
218
734237
4548
所以我们断开以太网的链接来创建一个空隙，
12:30
but again, like merely human hackers
219
738785
2668
但是，重申一遍，人类黑客都可以
12:33
routinely transgress air gaps
using social engineering.
using social engineering.
220
741453
3381
一次又一次以社会工程跨越这样的空隙。
12:36
Right now, as I speak,
221
744834
1259
现在，在我说话的时候，
12:38
I'm sure there is some employee
out there somewhere
out there somewhere
222
746093
2389
我肯定在这边的某个雇员，
12:40
who has been talked into handing out
her account details
her account details
223
748482
3346
曾近被要求交出他的账户明细，
12:43
by somebody claiming to be
from the I.T. department.
from the I.T. department.
224
751828
2746
给一个自称是电脑信息部门的人。
12:46
More creative scenarios are also possible,
225
754574
2127
其他的情况也有可能，
12:48
like if you're the A.I.,
226
756701
1315
比如如果你是人工智能，
12:50
you can imagine wiggling electrodes
around in your internal circuitry
around in your internal circuitry
227
758016
3532
你可以想象你用在你的体内
环绕复杂缠绕的电极
环绕复杂缠绕的电极
12:53
to create radio waves that you
can use to communicate.
can use to communicate.
228
761548
3462
创造出一种无线电波来交流。
12:57
Or maybe you could pretend to malfunction,
229
765010
2424
或者也许你可以假装你出了问题，
12:59
and then when the programmers open
you up to see what went wrong with you,
you up to see what went wrong with you,
230
767434
3497
然后程序师就把你打开看看哪里出错了，
13:02
they look at the source code -- Bam! --
231
770931
1936
他们找出了源代码——Bang！——
13:04
the manipulation can take place.
232
772867
2447
你就可以取得控制权了。
13:07
Or it could output the blueprint
to a really nifty technology,
to a really nifty technology,
233
775314
3430
或者它可以做出一个
非常漂亮的科技蓝图，
非常漂亮的科技蓝图，
13:10
and when we implement it,
234
778744
1398
当我们实现之后，
13:12
it has some surreptitious side effect
that the A.I. had planned.
that the A.I. had planned.
235
780142
4397
它有一些被人工智能计划好的
秘密的副作用。
秘密的副作用。
13:16
The point here is that we should
not be confident in our ability
not be confident in our ability
236
784539
3463
所以我们不能
对我们能够永远控制
对我们能够永远控制
13:20
to keep a superintelligent genie
locked up in its bottle forever.
locked up in its bottle forever.
237
788002
3808
一个超级智能体的能力
表示过度自信。
表示过度自信。
13:23
Sooner or later, it will out.
238
791810
2254
在不久后，它会逃脱出来。
13:27
I believe that the answer here
is to figure out
is to figure out
239
795034
3103
我相信我们需要弄明白
13:30
how to create superintelligent A.I.
such that even if -- when -- it escapes,
such that even if -- when -- it escapes,
240
798137
5024
如何创造出超级人工智能体，哪怕它逃走了，
13:35
it is still safe because it is
fundamentally on our side
fundamentally on our side
241
803161
3277
它仍然是无害的，因为它是我们这一边的，
13:38
because it shares our values.
242
806438
1899
因为它有我们的价值观。
13:40
I see no way around
this difficult problem.
this difficult problem.
243
808337
3210
我认为这是个不可避免的问题。
13:44
Now, I'm actually fairly optimistic
that this problem can be solved.
that this problem can be solved.
244
812557
3834
现在，我对这个问题能否被解决保持乐观。
13:48
We wouldn't have to write down
a long list of everything we care about,
a long list of everything we care about,
245
816391
3903
我们不需要写下
所有我们在乎的事情，
所有我们在乎的事情，
13:52
or worse yet, spell it out
in some computer language
in some computer language
246
820294
3643
或者，更糟地，把这些事情变成计算机语言，
13:55
like C++ or Python,
247
823937
1454
C++ 或者 Python，
13:57
that would be a task beyond hopeless.
248
825391
2767
这是个不可能的任务。
14:00
Instead, we would create an A.I.
that uses its intelligence
that uses its intelligence
249
828158
4297
而是，我们会创造出一个人工智能机器人，
用它自己的智慧
用它自己的智慧
14:04
to learn what we value,
250
832455
2771
来学习我们的价值观，
14:07
and its motivation system is constructed
in such a way that it is motivated
in such a way that it is motivated
251
835226
5280
它的激励制度可以激励它
14:12
to pursue our values or to perform actions
that it predicts we would approve of.
that it predicts we would approve of.
252
840506
5232
来追求我们的价值观
或者去做我们会赞成的事情。
或者去做我们会赞成的事情。
14:17
We would thus leverage
its intelligence as much as possible
its intelligence as much as possible
253
845738
3414
我们会因此最大地提高它的智力，
14:21
to solve the problem of value-loading.
254
849152
2745
来解决富有价值的问题。
14:24
This can happen,
255
852727
1512
这是有可能的，
14:26
and the outcome could be
very good for humanity.
very good for humanity.
256
854239
3596
结果可以使人类非常受益。
14:29
But it doesn't happen automatically.
257
857835
3957
但它不是自动发生的。
14:33
The initial conditions
for the intelligence explosion
for the intelligence explosion
258
861792
2998
智慧大爆炸的初始条件
14:36
might need to be set up
in just the right way
in just the right way
259
864790
2863
需要被正确地建立起来，
14:39
if we are to have a controlled detonation.
260
867653
3530
如果我们想要一切在掌握之中。
14:43
The values that the A.I. has
need to match ours,
need to match ours,
261
871183
2618
人工智能的价值观
要和我们的价值观相辅相成，
要和我们的价值观相辅相成，
14:45
not just in the familiar context,
262
873801
1760
不只是在熟悉的情况下，
14:47
like where we can easily check
how the A.I. behaves,
how the A.I. behaves,
263
875561
2438
比如当我们能很容易检查它的行为的时候，
14:49
but also in all novel contexts
that the A.I. might encounter
that the A.I. might encounter
264
877999
3234
但也要在所有人工智能可能会遇到的
前所未有的情况下，
前所未有的情况下，
14:53
in the indefinite future.
265
881233
1557
在没有界限的未来，
与我们的价值观相辅相成。
与我们的价值观相辅相成。
14:54
And there are also some esoteric issues
that would need to be solved, sorted out:
that would need to be solved, sorted out:
266
882790
4737
也有很多深奥的问题需要被分拣解决：
14:59
the exact details of its decision theory,
267
887527
2089
它如何做决定，
15:01
how to deal with logical
uncertainty and so forth.
uncertainty and so forth.
268
889616
2864
如何解决逻辑不确定性和类似的情况。
15:05
So the technical problems that need
to be solved to make this work
to be solved to make this work
269
893330
3102
所以技术上的待解决问题让这个任务
15:08
look quite difficult --
270
896432
1113
看起来有些困难：
15:09
not as difficult as making
a superintelligent A.I.,
a superintelligent A.I.,
271
897545
3380
并没有像做出一个超级智慧体一样困难，
15:12
but fairly difficult.
272
900925
2868
但是还是很难。
15:15
Here is the worry:
273
903793
1695
这使我们所担心的：
15:17
Making superintelligent A.I.
is a really hard challenge.
is a really hard challenge.
274
905488
4684
创造出一个超级智慧体确实是个很大的挑战。
15:22
Making superintelligent A.I. that is safe
275
910172
2548
创造出一个安全的超级智慧体，
15:24
involves some additional
challenge on top of that.
challenge on top of that.
276
912720
2416
是个更大的挑战。
15:28
The risk is that if somebody figures out
how to crack the first challenge
how to crack the first challenge
277
916216
3487
风险是，如果有人有办法解决第一个难题，
15:31
without also having cracked
the additional challenge
the additional challenge
278
919703
3001
却无法解决第二个
15:34
of ensuring perfect safety.
279
922704
1901
确保安全性的挑战。
15:37
So I think that we should
work out a solution
work out a solution
280
925375
3331
所以我认为我们应该预先想出
15:40
to the control problem in advance,
281
928706
2822
“控制性”的解决方法，
15:43
so that we have it available
by the time it is needed.
by the time it is needed.
282
931528
2660
这样我们就能在需要的时候用到它了。
15:46
Now it might be that we cannot solve
the entire control problem in advance
the entire control problem in advance
283
934768
3507
现在也许我们并不能
预先解决全部的控制性问题，
预先解决全部的控制性问题，
15:50
because maybe some elements
can only be put in place
can only be put in place
284
938275
3024
因为有些因素需要你了解
15:53
once you know the details of the
architecture where it will be implemented.
architecture where it will be implemented.
285
941299
3997
你要应用到的那个构架的细节才能实施。
15:57
But the more of the control problem
that we solve in advance,
that we solve in advance,
286
945296
3380
但如果我们能解决更多控制性的难题，
16:00
the better the odds that the transition
to the machine intelligence era
to the machine intelligence era
287
948676
4090
当我们迈入机器智能时代后
16:04
will go well.
288
952766
1540
就能更加顺利。
16:06
This to me looks like a thing
that is well worth doing
that is well worth doing
289
954306
4644
这对于我来说是个值得一试的东西，
16:10
and I can imagine that if
things turn out okay,
things turn out okay,
290
958950
3332
而且我能想象，如果一切顺利，
16:14
that people a million years from now
look back at this century
look back at this century
291
962282
4658
几百万年后的人类回首我们这个世纪，
16:18
and it might well be that they say that
the one thing we did that really mattered
the one thing we did that really mattered
292
966940
4002
他们也许会说，
我们所做的最最重要的事情，
我们所做的最最重要的事情，
16:22
was to get this thing right.
293
970942
1567
就是做了这个正确的决定。
16:24
Thank you.
294
972509
1689
谢谢。
16:26
(Applause)
295
974198
2813
（掌声）
ABOUT THE SPEAKER
Nick Bostrom - PhilosopherNick Bostrom asks big questions: What should we do, as individuals and as a species, to optimize our long-term prospects? Will humanity’s technological advancements ultimately destroy us?
Why you should listen
Philosopher Nick Bostrom envisioned a future full of human enhancement, nanotechnology and machine intelligence long before they became mainstream concerns. From his famous simulation argument -- which identified some striking implications of rejecting the Matrix-like idea that humans are living in a computer simulation -- to his work on existential risk, Bostrom approaches both the inevitable and the speculative using the tools of philosophy, probability theory, and scientific analysis.
Since 2005, Bostrom has led the Future of Humanity Institute, a research group of mathematicians, philosophers and scientists at Oxford University tasked with investigating the big picture for the human condition and its future. He has been referred to as one of the most important thinkers of our age.
Nick was honored as one of Foreign Policy's 2015 Global Thinkers .
His recent book Superintelligence advances the ominous idea that “the first ultraintelligent machine is the last invention that man need ever make.”
Nick Bostrom | Speaker | TED.com