The nightmare videos of childrens' YouTube — and what's wrong with the internet today | James Bridle

5,908,126 views

2018-07-13 ・ TED


New videos

The nightmare videos of childrens' YouTube — and what's wrong with the internet today | James Bridle

5,908,126 views ・ 2018-07-13

TED


请双击下面的英文字幕来播放视频。

翻译人员: jacks peng 校对人员: Joey Chen
00:12
I'm James.
0
12777
1183
我是詹姆斯。
00:13
I'm a writer and artist,
1
13984
1686
我是个作家和艺术家,
00:15
and I make work about technology.
2
15694
2341
我的工作跟技术有关。
00:18
I do things like draw life-size outlines of military drones
3
18454
3911
我做一些诸如绘制真实大小的军用无人机
00:22
in city streets around the world,
4
22389
1828
在世界各地城市街道上,
00:24
so that people can start to think and get their heads around
5
24241
2950
这样人们就可以去想象和思考
00:27
these really quite hard-to-see and hard-to-think-about technologies.
6
27215
3440
这些平时蛮难见到和蛮难想象的技术。
00:31
I make things like neural networks that predict the results of elections
7
31494
3836
我会制作一些神经网络的东西, 通过天气预报
00:35
based on weather reports,
8
35354
1737
来预测选举结果,
00:37
because I'm intrigued about
9
37115
1314
因为我感兴趣
00:38
what the actual possibilities of these weird new technologies are.
10
38453
3924
这些奇怪的新技术 实际的可能性是什么。
00:43
Last year, I built my own self-driving car.
11
43405
2426
去年,我还建造了自己的自动驾驶汽车。
00:45
But because I don't really trust technology,
12
45855
2526
但因为我不完全相信技术,
00:48
I also designed a trap for it.
13
48405
1928
我同时也给它设计了个包围圈。
00:50
(Laughter)
14
50777
1086
(笑声)
00:51
And I do these things mostly because I find them completely fascinating,
15
51887
4298
我做这些事情主要是 因为我发现他们真的很吸引人,
00:56
but also because I think when we talk about technology,
16
56209
2602
但也因为我觉得当我们谈论科技时,
00:58
we're largely talking about ourselves
17
58835
2619
我们其实主要是谈论我们自身
01:01
and the way that we understand the world.
18
61478
2299
以及我们理解世界的方式。
01:03
So here's a story about technology.
19
63801
2442
下面的故事就是有关科技的。
01:07
This is a "surprise egg" video.
20
67520
2830
这是“惊喜蛋”视频。
01:10
It's basically a video of someone opening up loads of chocolate eggs
21
70374
3348
内容就是人打开各种巧克力蛋
01:13
and showing the toys inside to the viewer.
22
73746
2126
向观众展示里面的玩具是啥。
01:16
That's it. That's all it does for seven long minutes.
23
76461
2649
就这样,这视频长达7分钟。
01:19
And I want you to notice two things about this.
24
79428
3051
我想让你们注意这两点:
01:22
First of all, this video has 30 million views.
25
82503
4074
首先,这视频有3千万观看量。
01:26
(Laughter)
26
86601
1275
(笑声)
01:28
And the other thing is,
27
88376
1166
另一件事是,
01:29
it comes from a channel that has 6.3 million subscribers,
28
89566
3869
它来自一个拥有630万订阅用户的频道,
01:33
that has a total of eight billion views,
29
93459
2680
该频道总播放量为80亿次,
01:36
and it's all just more videos like this --
30
96163
3106
这个频道主要都是这类的内容。
01:40
30 million people watching a guy opening up these eggs.
31
100256
3908
3千万人观看打开这些蛋蛋。
01:44
It sounds pretty weird, but if you search for "surprise eggs" on YouTube,
32
104188
4481
这听起来相当古怪, 但如果你在YouTube上搜索“惊喜蛋”,
01:48
it'll tell you there's 10 million of these videos,
33
108693
3523
它会告诉你这类视频多达1千万个,
01:52
and I think that's an undercount.
34
112240
1657
我认为这数字还低估了。
01:53
I think there's way, way more of these.
35
113921
1897
我认为数量远远大于此。
01:55
If you keep searching, they're endless.
36
115842
2242
如果你继续搜索,几乎无穷无尽。
01:58
There's millions and millions of these videos
37
118108
2159
这类视频,数不胜数。
02:00
in increasingly baroque combinations of brands and materials,
38
120291
3454
如巴洛克般花哨地 混合着各种品牌和材料,
02:03
and there's more and more of them being uploaded every single day.
39
123769
3846
每天这类视频上传的数量还越来越多。
02:07
Like, this is a strange world. Right?
40
127639
3511
这真是个奇怪的世界,对吗?
02:11
But the thing is, it's not adults who are watching these videos.
41
131174
3383
但事实是,不是成年人在看这些视频。
02:14
It's kids, small children.
42
134581
2921
是小朋友,小孩子们。
02:17
These videos are like crack for little kids.
43
137526
2154
这些视频就像小孩子的毒品。
02:19
There's something about the repetition,
44
139704
2075
是那种让人反复的东西,
02:21
the constant little dopamine hit of the reveal,
45
141803
2468
小多巴胺不断涌现出来,
02:24
that completely hooks them in.
46
144295
1866
完全令他们着迷不已。
02:26
And little kids watch these videos over and over and over again,
47
146185
4809
孩子们一遍又一遍地看这些视频,
02:31
and they do it for hours and hours and hours.
48
151018
2327
一个小时接一个小时地看。
02:33
And if you try and take the screen away from them,
49
153369
2356
如果你试图把屏幕关上,
02:35
they'll scream and scream and scream.
50
155749
1782
他们会朝你尖叫,尖叫,尖叫
02:37
If you don't believe me --
51
157555
1262
假如你不相信我——
02:38
and I've already seen people in the audience nodding --
52
158841
2607
我已经看到在座各位已经有人点头——
02:41
if you don't believe me, find someone with small children and ask them,
53
161472
3391
如果你不相信我,找个有小孩的人问问,
02:44
and they'll know about the surprise egg videos.
54
164887
2340
他们肯定知道这些惊喜蛋视频。
02:47
So this is where we start.
55
167251
2070
这就是我们开始的地方。
02:49
It's 2018, and someone, or lots of people,
56
169345
3642
2018年,有人,或很多人
02:53
are using the same mechanism that, like, Facebook and Instagram are using
57
173011
3941
在使用类似的机制,如Facebook 和Instagram就在使用的机制
02:56
to get you to keep checking that app,
58
176976
1989
来让你沉迷他们的应用,
02:58
and they're using it on YouTube to hack the brains of very small children
59
178989
3985
他们在YouTube上用这些机制 劫持小孩子的脑袋
03:02
in return for advertising revenue.
60
182998
1958
来换取广告收入。
03:06
At least, I hope that's what they're doing.
61
186346
2001
至少,我希望这是他们在做的事情。
03:08
I hope that's what they're doing it for,
62
188371
1955
我希望这是他们做这些事情的目的。
03:10
because there's easier ways of making ad revenue on YouTube.
63
190350
5308
因为Youtube上有更简单的 赚取广告收入的方式
03:15
You can just make stuff up or steal stuff.
64
195682
2332
你可以捏造或者干脆照搬其他人的作品。
03:18
So if you search for really popular kids' cartoons
65
198038
2635
所以如果你搜索真正流行的儿童卡通
03:20
like "Peppa Pig" or "Paw Patrol,"
66
200697
1654
比如“小猪佩奇”或者“狗狗巡逻队”
03:22
you'll find there's millions and millions of these online as well.
67
202375
3147
你会发现这类视频也有不计其数。
03:25
Of course, most of them aren't posted by the original content creators.
68
205546
3352
当然,其中绝大部分并非由 内容版权方上传。
03:28
They come from loads and loads of different random accounts,
69
208922
2999
他们来自大量不同的随机账号,
03:31
and it's impossible to know who's posting them
70
211945
2240
几乎无法知道谁在上传,
03:34
or what their motives might be.
71
214209
1822
他们的动机是什么 。
03:36
Does that sound kind of familiar?
72
216428
1930
这是不是听起来有点熟悉?
03:38
Because it's exactly the same mechanism
73
218382
1980
因为这完全就是同一套机制,
03:40
that's happening across most of our digital services,
74
220386
2600
几乎每个网络平台上都会利用它,
03:43
where it's impossible to know where this information is coming from.
75
223010
3207
我们根本无法知道这些信息的来源。
03:46
It's basically fake news for kids,
76
226241
1829
这就是给儿童看的假新闻,
03:48
and we're training them from birth
77
228094
2161
我们从他们出生开始
03:50
to click on the very first link that comes along,
78
230279
2506
训练他们点击每一个链接,
03:52
regardless of what the source is.
79
232809
1953
不管信息来源何处。
03:54
That's doesn't seem like a terribly good idea.
80
234786
2603
这听起来可不是个好主意。
03:58
Here's another thing that's really big on kids' YouTube.
81
238399
2710
这个视频在儿童的YouTube频道 也相当流行。
04:01
This is called the "Finger Family Song."
82
241133
1928
这叫做“手指之歌”。
04:03
I just heard someone groan in the audience.
83
243085
2018
我刚听到有人在观众席上叹气。
04:05
This is the "Finger Family Song."
84
245127
1624
这是“手指之歌”。
04:06
This is the very first one I could find.
85
246775
1930
这是我能找到的最初版本。
04:08
It's from 2007, and it only has 200,000 views,
86
248729
2829
来自2007年,当时只有20万播放量,
04:11
which is, like, nothing in this game.
87
251582
1976
在这场游戏中根本微不足道。
04:13
But it has this insanely earwormy tune,
88
253582
2852
但它拥有不可思议的绕梁三日音调,
04:16
which I'm not going to play to you,
89
256458
1682
我可不打算播放给你们听,
04:18
because it will sear itself into your brain
90
258164
2008
因为它也会钻进你的脑袋里,
04:20
in the same way that it seared itself into mine,
91
260196
2395
就跟它钻进我脑袋一样,
04:22
and I'm not going to do that to you.
92
262615
1770
我不打算那样对你们。
04:24
But like the surprise eggs,
93
264409
1344
但跟惊喜蛋一样
04:25
it's got inside kids' heads
94
265777
2164
它留在了孩子的头脑里
04:27
and addicted them to it.
95
267965
1607
让他们产生沉迷。
04:29
So within a few years, these finger family videos
96
269596
2531
所以不到几年,这些手指之歌视频
04:32
start appearing everywhere,
97
272151
1303
开始无处不在,
04:33
and you get versions in different languages
98
273478
2029
于是就有了各种语言的版本,
04:35
with popular kids' cartoons using food
99
275531
2121
有流行卡通使用食物
04:37
or, frankly, using whatever kind of animation elements
100
277676
2550
或者使用任何动画元素的版本。
04:40
you seem to have lying around.
101
280250
2252
你就像躺在上面一样。
04:43
And once again, there are millions and millions and millions of these videos
102
283002
5197
再一次,这些不计其数视频
04:48
available online in all of these kind of insane combinations.
103
288223
3435
以各种疯狂的组合方式在网上出现。
04:51
And the more time you start to spend with them,
104
291682
2228
你花在这些上面的时间越多,
04:53
the crazier and crazier you start to feel that you might be.
105
293934
3694
你就会觉得自己越来越疯癫。
04:57
And that's where I kind of launched into this,
106
297652
3333
这就是我要开始讲的,
05:01
that feeling of deep strangeness and deep lack of understanding
107
301009
3666
始终,我有一种陌生感, 我也不理解
05:04
of how this thing was constructed that seems to be presented around me.
108
304699
4175
这些东西如何被构建出来。
05:08
Because it's impossible to know where these things are coming from.
109
308898
3167
因为我没法知道这些东西来自哪里。
05:12
Like, who is making them?
110
312089
1241
比如,谁制作了它们?
05:13
Some of them appear to be made of teams of professional animators.
111
313354
3143
有些视频似乎来自专业动画团队。
05:16
Some of them are just randomly assembled by software.
112
316521
2882
有些只是随机由软件合成。
05:19
Some of them are quite wholesome-looking young kids' entertainers.
113
319427
4253
有些视频中有看起来友好的表演者
05:23
And some of them are from people
114
323704
1552
但也有一些视频中的人
05:25
who really clearly shouldn't be around children at all.
115
325280
3007
一看就是儿童不宜的。
05:28
(Laughter)
116
328311
1615
(笑声)
05:30
And once again, this impossibility of figuring out who's making this stuff --
117
330987
4640
再一次,几乎不可能搞清楚是谁 制作了这些东西。
05:35
like, this is a bot?
118
335651
1156
是机器人?
05:36
Is this a person? Is this a troll?
119
336831
2647
是人?还是网络喷子?
05:39
What does it mean that we can't tell the difference
120
339502
2382
我们无法分辨出彼此差别
05:41
between these things anymore?
121
341908
1583
到底意味着什么?
05:43
And again, doesn't that uncertainty feel kind of familiar right now?
122
343515
4848
再一次,这种不确定性是不是有点熟悉?
05:50
So the main way people get views on their videos --
123
350145
2580
所以人们获得观看量的主要方法是
05:52
and remember, views mean money --
124
352749
1707
记住,观看量意味着金钱,
05:54
is that they stuff the titles of these videos with these popular terms.
125
354480
4742
是他们用热门词充斥这些视频的标题。
05:59
So you take, like, "surprise eggs"
126
359246
1687
以“惊喜蛋”为例,
06:00
and then you add "Paw Patrol," "Easter egg,"
127
360957
2066
你会增加“狗狗巡逻队”、“复活节彩蛋“
06:03
or whatever these things are,
128
363047
1393
或者任何其他词语,
06:04
all of these words from other popular videos into your title,
129
364464
2893
这些来自其他热门视频的词 添加到你的标题,
06:07
until you end up with this kind of meaningless mash of language
130
367381
2968
直到你最终得到这种对人类而言
06:10
that doesn't make sense to humans at all.
131
370373
2498
毫无意义的词语混杂。
06:12
Because of course it's only really tiny kids who are watching your video,
132
372895
3546
因为当然只有很小的孩子在看你的视频,
06:16
and what the hell do they know?
133
376465
1827
他们能知道什么?
06:18
Your real audience for this stuff is software.
134
378316
3007
你这些东西的真正观众是软件,
06:21
It's the algorithms.
135
381347
1156
是算法。
06:22
It's the software that YouTube uses
136
382527
1855
这是YouTube使用来
06:24
to select which videos are like other videos,
137
384406
2483
选择哪个视频像哪个视频,
06:26
to make them popular, to make them recommended.
138
386913
2243
让他们流行和推荐的算法。
06:29
And that's why you end up with this kind of completely meaningless mash,
139
389180
3461
所以你最终得到的就是 这种完全没有意义的大杂烩,
06:32
both of title and of content.
140
392665
2071
不管是标题还是内容。
06:35
But the thing is, you have to remember,
141
395792
1894
但事情是,你需要记住,
06:37
there really are still people within this algorithmically optimized system,
142
397710
4478
这个优化的算法系统还是需要人的参与,
06:42
people who are kind of increasingly forced to act out
143
402212
2790
这些人被迫面对处理
06:45
these increasingly bizarre combinations of words,
144
405026
3066
这些越来越奇怪的词语组合,
06:48
like a desperate improvisation artist responding to the combined screams
145
408116
5173
就像一个绝望的即兴艺术家 要对上百万尖叫孩子组合
06:53
of a million toddlers at once.
146
413313
2203
做出回应一样。
06:57
There are real people trapped within these systems,
147
417168
2468
一些人则被困在这个系统里面,
06:59
and that's the other deeply strange thing about this algorithmically driven culture,
148
419660
4055
另一个很奇怪的事情是 关于算法驱动文化,
07:03
because even if you're human,
149
423739
1381
因为即便你是人类,
07:05
you have to end up behaving like a machine
150
425144
2145
你最终也会变得像机器一样,
07:07
just to survive.
151
427313
1800
只是为了生存。
07:09
And also, on the other side of the screen,
152
429137
2100
而且,屏幕的另一面是,
07:11
there still are these little kids watching this stuff,
153
431261
2947
这些小孩仍然在看这些视频,
07:14
stuck, their full attention grabbed by these weird mechanisms.
154
434232
4206
他们的注意力完全被 这些奇怪的机制所左右。
07:18
And most of these kids are too small to even use a website.
155
438768
2798
大部分小孩年纪都很小, 甚至还不会使用网页。
07:21
They're just kind of hammering on the screen with their little hands.
156
441590
3276
他们只会用他们的小手敲打着屏幕。
07:24
And so there's autoplay,
157
444890
1217
这是自动播放按钮,
07:26
where it just keeps playing these videos over and over and over in a loop,
158
446131
3579
它会不断循环播放这些视频,
07:29
endlessly for hours and hours at a time.
159
449734
2059
无休止持续数小时。
07:31
And there's so much weirdness in the system now
160
451817
2843
现在这个系统里有很多奇怪的东西
07:34
that autoplay takes you to some pretty strange places.
161
454684
3009
自动播放会带你去一些非常奇怪的地方。
07:37
This is how, within a dozen steps,
162
457717
2488
这里演示的是,在十几个步骤里,
07:40
you can go from a cute video of a counting train
163
460229
3158
你可能会从数火车的有趣视频
07:43
to masturbating Mickey Mouse.
164
463411
2442
播到手淫的米老鼠。
07:46
Yeah. I'm sorry about that.
165
466529
2288
是的,非常抱歉。
07:48
This does get worse.
166
468841
1700
事情确实变糟糕了。
07:50
This is what happens
167
470565
1282
这就是会发生的事情:
07:51
when all of these different keywords,
168
471871
3086
当所有这些不同的关键词
07:54
all these different pieces of attention,
169
474981
2461
所有可以吸引注意力的内容
07:57
this desperate generation of content,
170
477466
2807
所有迫不及待地让视频播出的心理
08:00
all comes together into a single place.
171
480297
2582
都在同一个地方冒出来。
08:03
This is where all those deeply weird keywords come home to roost.
172
483871
4472
这是这些非常怪异的关键词 自食其果的地方。
08:08
You cross-breed the finger family video
173
488367
2391
你混合了手指家庭视频
08:10
with some live-action superhero stuff,
174
490782
2088
和一些超级英雄的真人动作视频,
08:12
you add in some weird, trollish in-jokes or something,
175
492894
3256
你加入一些奇怪,恶搞或其他元素。
08:16
and suddenly, you come to a very weird place indeed.
176
496174
3366
突然之间,你就会进入非常怪异的领域。
08:19
The stuff that tends to upset parents
177
499564
2113
那些会让父母感到不安的东西
08:21
is the stuff that has kind of violent or sexual content, right?
178
501701
3331
是充满暴力和色情的内容,对吧?
08:25
Children's cartoons getting assaulted,
179
505056
2822
儿童们的卡通形象被侵犯,
08:27
getting killed,
180
507902
2018
被杀害,
08:29
weird pranks that actually genuinely terrify children.
181
509944
3343
怪异的恶作剧,真的让孩子们感到恐惧。
08:33
What you have is software pulling in all of these different influences
182
513311
3675
现在你所能得到的是软件 把这些不同的内容
自动组合成了儿童们的噩梦。
08:37
to automatically generate kids' worst nightmares.
183
517010
2961
08:39
And this stuff really, really does affect small children.
184
519995
2701
这些东西真的,真的会影响小孩子。
08:42
Parents report their children being traumatized,
185
522720
2866
父母报告说他们的孩子受到了精神创伤,
08:45
becoming afraid of the dark,
186
525610
1392
变得害怕黑暗,
08:47
becoming afraid of their favorite cartoon characters.
187
527026
3050
开始害怕他们喜爱的卡通角色。
08:50
If you take one thing away from this, it's that if you have small children,
188
530524
3611
如果你可以从中学到一件事: 如果你有小孩,
08:54
keep them the hell away from YouTube.
189
534159
1996
让他们远离YouTube。
08:56
(Applause)
190
536743
3949
(掌声)
09:02
But the other thing, the thing that really gets to me about this,
191
542504
3096
但另一件事情, 这事情真正让我关心这个的是
09:05
is that I'm not sure we even really understand how we got to this point.
192
545624
4629
我不确定我们是否真正理解 事情是如何发展到这一步的。
09:10
We've taken all of this influence, all of these things,
193
550951
2931
我们已经看到了所有这些影响, 所有这些东西,
09:13
and munged them together in a way that no one really intended.
194
553906
2953
以一种没人意料得到的方式 组合在一起。
09:16
And yet, this is also the way that we're building the entire world.
195
556883
3156
然而,这也是我们构建 整个世界的方式。
09:20
We're taking all of this data,
196
560063
1773
我们得到了所有这些数据,
09:21
a lot of it bad data,
197
561860
1447
无数的坏数据,
09:23
a lot of historical data full of prejudice,
198
563331
3029
大量充满偏见的历史数据,
09:26
full of all of our worst impulses of history,
199
566384
2837
充斥着历史上最糟糕冲动的数据,
09:29
and we're building that into huge data sets
200
569245
2049
我们把他们制作成海量数据集
09:31
and then we're automating it.
201
571318
1423
然后让它们自动化。
09:32
And we're munging it together into things like credit reports,
202
572765
3502
我们把它们生成信用报告,
09:36
into insurance premiums,
203
576291
1634
保险费用,
09:37
into things like predictive policing systems,
204
577949
2693
预测性警务系统,
09:40
into sentencing guidelines.
205
580666
1762
量刑建议。
09:42
This is the way we're actually constructing the world today
206
582452
2821
这就是我们今天基于这些数据
09:45
out of this data.
207
585297
1151
构建世界的方式。
09:46
And I don't know what's worse,
208
586472
1698
我不知道哪个更糟糕:
09:48
that we built a system that seems to be entirely optimized
209
588194
3228
是我们似乎建造了一个完全
09:51
for the absolute worst aspects of human behavior,
210
591446
2808
适合人类负面行为的优化系统,
09:54
or that we seem to have done it by accident,
211
594278
2425
或者我们只是无意中造就了它,
09:56
without even realizing that we were doing it,
212
596727
2207
完全没有意识,
09:58
because we didn't really understand the systems that we were building,
213
598958
3382
因为我们并不理解我们们正创建的系统,
10:02
and we didn't really understand how to do anything differently with it.
214
602364
3683
我们也并不知道是否有 其他不同的方式来使用它。
10:06
There's a couple of things I think that really seem to be driving this
215
606769
3365
我认为有几个因素看起来导致了
10:10
most fully on YouTube,
216
610158
1189
YouTube内容事件的发生,
10:11
and the first of those is advertising,
217
611371
1827
其中之首是广告。
10:13
which is the monetization of attention
218
613222
2837
一种靠注意力赢利的模式
10:16
without any real other variables at work,
219
616083
3136
而不考虑其他变量因素,
10:19
any care for the people who are actually developing this content,
220
619243
3885
不关心是哪些人确实开发了这些内容。
10:23
the centralization of the power, the separation of those things.
221
623152
3636
权力的集中,东西的分散。
10:26
And I think however you feel about the use of advertising
222
626812
3144
我想,不管你对用广告来宣传某物
10:29
to kind of support stuff,
223
629980
1238
有着什么样的看法,
10:31
the sight of grown men in diapers rolling around in the sand
224
631242
3067
在沙里翻滚的穿纸尿裤的大人
10:34
in the hope that an algorithm that they don't really understand
225
634333
2983
希望无知的机器算法
10:37
will give them money for it
226
637340
1315
会发钱给他们。
10:38
suggests that this probably isn't the thing
227
638679
2037
意味着这类事情
10:40
that we should be basing our society and culture upon,
228
640740
2563
不应该是我们社会和文化所依托的
10:43
and the way in which we should be funding it.
229
643327
2160
也不该是我们应该资助的。
10:45
And the other thing that's kind of the major driver of this is automation,
230
645511
3519
另一个主要驱动因素是自动化,
10:49
which is the deployment of all of this technology
231
649054
2329
是所有这些技术的部署
10:51
as soon as it arrives, without any kind of oversight,
232
651407
2521
多多益善,不加审察。
10:53
and then once it's out there,
233
653952
1412
而一旦出了问题,
10:55
kind of throwing up our hands and going, "Hey, it's not us, it's the technology."
234
655388
3843
就两手一摊,“嗨, 不是我们做的,技术搞的。”
10:59
Like, "We're not involved in it."
235
659255
1642
又如,“我们没有参与其中。”
11:00
That's not really good enough,
236
660921
1767
这还不够好,
11:02
because this stuff isn't just algorithmically governed,
237
662712
2710
因为这些平台不仅由算法控制,
11:05
it's also algorithmically policed.
238
665446
2498
还有被算法监管。
11:07
When YouTube first started to pay attention to this,
239
667968
2848
当YouTube开始关注到这个问题时,
11:10
the first thing they said they'd do about it
240
670840
2087
他们说他们会做的首要事情是
11:12
was that they'd deploy better machine learning algorithms
241
672951
2695
部署更好的机器学习算法
11:15
to moderate the content.
242
675670
1329
来控制内容。
11:17
Well, machine learning, as any expert in it will tell you,
243
677023
3485
好吧,机器学习, 很多专家都会告诉你,
11:20
is basically what we've started to call
244
680532
1896
就是那种我们开始说的
11:22
software that we don't really understand how it works.
245
682452
2588
我们无法真正理解它如何工作的软件。
11:25
And I think we have enough of that already.
246
685064
3983
我认为我们已经受够它了。
11:29
We shouldn't be leaving this stuff up to AI to decide
247
689071
3166
我们不该让AI决定
11:32
what's appropriate or not,
248
692261
1251
什么是合适与否的内容,
11:33
because we know what happens.
249
693536
1436
因为我们知道会发生什么。
11:34
It'll start censoring other things.
250
694996
1688
它会开始审查其他事情。
11:36
It'll start censoring queer content.
251
696708
1783
它会开始审查同性内容。
11:38
It'll start censoring legitimate public speech.
252
698515
2237
它会开始审查合法的公开演讲。
11:40
What's allowed in these discourses,
253
700776
1925
在演讲中应该允许什么,
11:42
it shouldn't be something that's left up to unaccountable systems.
254
702725
3097
这不该是留给不可靠系统去决定的事。
11:45
It's part of a discussion all of us should be having.
255
705846
2947
这是我们所有人都应该讨论的事情。
11:48
But I'd leave a reminder
256
708817
1308
但我要留个提醒
11:50
that the alternative isn't very pleasant, either.
257
710149
2753
替代方案也并非完美。
11:52
YouTube also announced recently
258
712926
1535
YouTube最近也宣布
11:54
that they're going to release a version of their kids' app
259
714485
2767
它们会推出它们的儿童版APP
11:57
that would be entirely moderated by humans.
260
717276
2407
将会完全由人工监管。
12:00
Facebook -- Zuckerberg said much the same thing at Congress,
261
720134
3618
Facebook——扎克伯格在国会上 也说了类似的话,
12:03
when pressed about how they were going to moderate their stuff.
262
723776
2987
当被问到如何去调整他们的产品时。
12:06
He said they'd have humans doing it.
263
726787
1747
他说他们会让人处理。
12:08
And what that really means is,
264
728558
1459
那其实隐含的意思是,
12:10
instead of having toddlers being the first person to see this stuff,
265
730041
3223
以其让蹒跚学步的孩子 成为第一个看到这些东西的人,
12:13
you're going to have underpaid, precarious contract workers
266
733288
2788
你打算让那些工资过低,不稳定,
12:16
without proper mental health support
267
736100
1726
没有心理辅导的临时工
12:17
being damaged by it as well.
268
737850
1376
被那些视频伤害。
12:19
(Laughter)
269
739250
1096
(笑声)
12:20
And I think we can all do quite a lot better than that.
270
740370
2601
我想我们都可以做得比这更好。
12:22
(Applause)
271
742995
2499
(鼓掌)
12:26
The thought, I think, that brings those two things together, really, for me,
272
746068
4613
总结这两件事,我的想法在于:
12:30
is agency.
273
750705
1420
自主能动性
12:32
It's like, how much do we really understand -- by agency, I mean:
274
752149
3157
就像,我们是否真正懂得“自主能动性”:
12:35
how we know how to act in our own best interests.
275
755330
4390
我们是否知怎样按照 自己的最佳利益行事。
12:39
Which -- it's almost impossible to do
276
759744
1787
而这几乎无法在
12:41
in these systems that we don't really fully understand.
277
761555
3485
我们并不完全理解的系统中实现。
12:45
Inequality of power always leads to violence.
278
765064
3071
权力的不对等总会导致暴力。
12:48
And we can see inside these systems
279
768159
1685
我们可以在这些系统中看到
12:49
that inequality of understanding does the same thing.
280
769868
2611
理解的不对等也会造成同样的结果。
12:52
If there's one thing that we can do to start to improve these systems,
281
772503
3779
如果我们能够做一件事情 去提升这些系统,
12:56
it's to make them more legible to the people who use them,
282
776306
2718
那就是让它们变得更透明
12:59
so that all of us have a common understanding
283
779048
2196
这样我们所有的人都对其中的情况
13:01
of what's actually going on here.
284
781268
1851
有一个共同的理解。
13:03
The thing, though, I think most about these systems
285
783970
2968
但是,我认为这些系统的关键问题
13:06
is that this isn't, as I hope I've explained, really about YouTube.
286
786962
3857
我希望我已解释过了, 这真的并不是YouTube的问题。
13:10
It's about everything.
287
790843
1312
任何事都是一样。
13:12
These issues of accountability and agency,
288
792179
2444
这些问责和自主能动性,
13:14
of opacity and complexity,
289
794647
2225
不透明性和复杂性问题,
13:16
of the violence and exploitation that inherently results
290
796896
3177
暴力和剥削问题
是因为权力集中于少数人手中。
13:20
from the concentration of power in a few hands --
291
800097
2794
13:22
these are much, much larger issues.
292
802915
2579
这些都是更大的问题。
13:26
And they're issues not just of YouTube and not just of technology in general,
293
806395
3687
他们不仅是YouTube的问题, 而且不仅仅是科技问题,
13:30
and they're not even new.
294
810106
1265
它们甚至都不是新问题。
13:31
They've been with us for ages.
295
811395
1461
他们已经存在很久了。
13:32
But we finally built this system, this global system, the internet,
296
812880
4390
但是我们最终建立了这个系统, 这个全球系统,互联网,
13:37
that's actually showing them to us in this extraordinary way,
297
817294
3019
它实际上是用这种特殊的方式 向我们展示,
13:40
making them undeniable.
298
820337
1547
他们至高无上。
13:41
Technology has this extraordinary capacity
299
821908
2820
技术有这种非凡的能力
13:44
to both instantiate and continue
300
824752
3973
去具现化和继续
13:48
all of our most extraordinary, often hidden desires and biases
301
828749
4248
我们所有最卓越,而通常 被隐藏的欲望和偏见,
13:53
and encoding them into the world,
302
833021
1866
并把它们编码到世界中,
13:54
but it also writes them down so that we can see them,
303
834911
3474
但是它也会把它们写下来, 这样我们就能看到它们了。
13:58
so that we can't pretend they don't exist anymore.
304
838409
3330
所以我们不能假装这些问题不再存在了。
14:01
We need to stop thinking about technology as a solution to all of our problems,
305
841763
4319
我们需要停止把技术 当作是解决一切问题的良方,
14:06
but think of it as a guide to what those problems actually are,
306
846106
3757
而应把它看作指引我们发现 问题的指南针,
14:09
so we can start thinking about them properly
307
849887
2144
这样我们才可以开始正视它们
14:12
and start to address them.
308
852055
1766
并开始着手解决它们。
14:13
Thank you very much.
309
853845
1335
谢谢!
14:15
(Applause)
310
855204
5192
(掌声)
14:21
Thank you.
311
861733
1188
谢谢!
14:22
(Applause)
312
862945
2869
(掌声)
14:28
Helen Walters: James, thank you for coming and giving us that talk.
313
868839
3178
海伦·沃尔特斯:詹姆斯, 谢谢你的到来和这个演讲。
14:32
So it's interesting:
314
872041
1189
这很有趣:
14:33
when you think about the films where the robotic overlords take over,
315
873254
3495
当在电影中,当那些机器人 开始统治世界时,
14:36
it's all a bit more glamorous than what you're describing.
316
876773
3279
那个景象好似比你描述的更加宏伟。
14:40
But I wonder -- in those films, you have the resistance mounting.
317
880076
3749
但我想知道的是,在那些电影里, 往往会有人类抵抗军
14:43
Is there a resistance mounting towards this stuff?
318
883849
3216
现实中是否也存在这些反抗军呢?
14:47
Do you see any positive signs, green shoots of resistance?
319
887089
3796
你有没有看到任何积极的迹象, 抵抗的萌芽?
14:52
James Bridle: I don't know about direct resistance,
320
892507
2416
詹姆斯·布里德尔:我不清楚 直接的抵抗力量,
14:54
because I think this stuff is super long-term.
321
894947
2264
因为我觉得这是非常长期的问题。
14:57
I think it's baked into culture in really deep ways.
322
897235
2510
我认为它在文化中根深蒂固。
14:59
A friend of mine, Eleanor Saitta, always says
323
899769
2132
我的朋友,埃莉诺·萨塔,总是说
15:01
that any technological problems of sufficient scale and scope
324
901935
3609
任何足够规模和范围的科技问题
15:05
are political problems first of all.
325
905568
2267
首先是政治问题。
15:07
So all of these things we're working to address within this
326
907859
2785
所有这些我们在努力解决的问题
15:10
are not going to be addressed just by building the technology better,
327
910668
3274
不能仅仅通过研发更好的技术来解决,
15:13
but actually by changing the society that's producing these technologies.
328
913966
3464
而应通过改变产生 这些技术的社会来解决。
15:17
So no, right now, I think we've got a hell of a long way to go.
329
917454
3027
所以,现在我想我们还有很长的路要走。
15:20
But as I said, I think by unpacking them,
330
920505
1986
但就像我说的,我想把它们拆解,
15:22
by explaining them, by talking about them super honestly,
331
922515
2697
通过解构他们,诚实地谈论他们
15:25
we can actually start to at least begin that process.
332
925236
2505
我们至少可以开始这个过程。
15:27
HW: And so when you talk about legibility and digital literacy,
333
927765
3562
沃尔特斯: 还有当你谈到易读性和 数字素养时,
15:31
I find it difficult to imagine
334
931351
1591
我觉得非常难以想象
15:32
that we need to place the burden of digital literacy on users themselves.
335
932966
3680
我们要把数字扫盲的负担 放在用户身上。
15:36
But whose responsibility is education in this new world?
336
936670
4562
但在这个新世界,教育是谁的责任?
15:41
JB: Again, I think this responsibility is kind of up to all of us,
337
941256
3612
布里德尔:再次,我认为这个责任是 我们所有人的责任,
15:44
that everything we do, everything we build, everything we make,
338
944892
2984
我们所做的、所构建的、所制造的一切,
15:47
needs to be made in a consensual discussion
339
947900
3692
需要与每个回避问题的人
15:51
with everyone who's avoiding it;
340
951616
1940
以相互求同为目的,进行讨论;
15:53
that we're not building systems intended to trick and surprise people
341
953580
4341
我们建造这些系统并不是为了 去诱惑人
15:57
into doing the right thing,
342
957945
2300
去做正确的事情。
16:00
but that they're actually involved in every step in educating them,
343
960269
3236
而是教育他们的每一步,
16:03
because each of these systems is educational.
344
963529
2278
因为每一个系统都是有教育意义的。
16:05
That's what I'm hopeful about, about even this really grim stuff,
345
965831
3102
这就是我所希望的, 即使是非常严峻的问题,
16:08
that if you can take it and look at it properly,
346
968957
2262
如果你能正确地看待它,
16:11
it's actually in itself a piece of education
347
971243
2089
这本身就是一种教育,
16:13
that allows you to start seeing how complex systems come together and work
348
973356
3762
让你可以开始看到复杂的系统 是如何结合在一起工作的
16:17
and maybe be able to apply that knowledge elsewhere in the world.
349
977142
3501
也许能够将这些知识 应用到世界的其他地方。
16:20
HW: James, it's such an important discussion,
350
980667
2115
沃尔特斯:詹姆斯, 这是一个很重要的讨论,
16:22
and I know many people here are really open and prepared to have it,
351
982806
3227
我知道这里很多人都乐于和你讨论,
16:26
so thanks for starting off our morning.
352
986057
1859
谢谢你今早的开场演讲。
16:27
JB: Thanks very much. Cheers.
353
987940
1400
布里德尔:非常感谢。
16:29
(Applause)
354
989364
1651
(掌声)
关于本网站

这个网站将向你介绍对学习英语有用的YouTube视频。你将看到来自世界各地的一流教师教授的英语课程。双击每个视频页面上显示的英文字幕,即可从那里播放视频。字幕会随着视频的播放而同步滚动。如果你有任何意见或要求,请使用此联系表与我们联系。

https://forms.gle/WvT1wiN1qDtmnspy7