How do we learn to work with intelligent machines? | Matt Beane

64,101 views ・ 2019-02-21

TED


请双击下面的英文字幕来播放视频。

翻译人员: Ruijie Wu 校对人员:
00:13
It’s 6:30 in the morning,
0
13292
1875
清晨六点半,
00:15
and Kristen is wheeling her prostate patient into the OR.
1
15583
4875
克里斯汀正推着 她的前列腺病人进手术室。
00:21
She's a resident, a surgeon in training.
2
21500
2250
她是一名实习住院外科医生,
00:24
It’s her job to learn.
3
24333
2167
学习是她工作的一部分。
00:27
Today, she’s really hoping to do some of the nerve-sparing,
4
27292
3351
这天,她非常想 参与进行神经保留手术,
00:30
extremely delicate dissection that can preserve erectile function.
5
30667
3875
这要求医生有极度精细的切割技巧, 以让病人恢复勃起的功能。
00:35
That'll be up to the attending surgeon, though, but he's not there yet.
6
35500
3338
不过,这还要看主治医生的意思, 但那会儿他并不在手术室。
00:39
She and the team put the patient under,
7
39625
2393
克里斯汀和其他手术人员 给病人打了麻醉。
首先,她在病人的下腹部 切开了一道8英寸的切口,
00:42
and she leads the initial eight-inch incision in the lower abdomen.
8
42042
3708
00:47
Once she’s got that clamped back, she tells the nurse to call the attending.
9
47042
3586
当她把切口固定好, 便让护士打电话给主治医生。
00:51
He arrives, gowns up,
10
51583
2292
主治医生赶到后,穿上手术服。
00:54
And from there on in, their four hands are mostly in that patient --
11
54458
5792
接着,两人共同开始手术, 他们四只手都在病人体内,
01:00
with him guiding but Kristin leading the way.
12
60708
2917
主治医生负责指导, 克里斯汀则主导了手术。
01:04
When the prostates out (and, yes, he let Kristen do a little nerve sparing),
13
64875
4643
当病人的前列腺被取出后,主治医生 让她进行了部分神经保留的操作,
01:09
he rips off his scrubs.
14
69542
1226
他脱掉了手术服,
01:10
He starts to do paperwork.
15
70792
1375
开始处理一些文件。
01:12
Kristen closes the patient by 8:15,
16
72833
5375
而克里斯汀在一个 初级住院医生的协助下
01:18
with a junior resident looking over her shoulder.
17
78583
2435
于8:15完成了手术,
克里斯汀还让他 给病人做了最后的缝合。
01:21
And she lets him do the final line of sutures.
18
81042
3083
01:24
Kristen feels great.
19
84833
3042
克里斯汀感觉好极了,
01:28
Patient’s going to be fine,
20
88250
1559
病人很快就会恢复,
01:29
and no doubt she’s a better surgeon than she was at 6:30.
21
89833
3167
而她也无疑比凌晨六点半时的 自己更进了一步。
01:34
Now this is extreme work.
22
94208
2834
虽然,医生的工作挑战性极高。
01:37
But Kristin’s learning to do her job the way that most of us do:
23
97417
3833
但克里斯汀的学习过程 其实和我们的并无分别,
01:41
watching an expert for a bit,
24
101625
1893
通过观察专家的操作,
01:43
getting involved in easy, safe parts of the work
25
103542
3142
从简单、安全的部分开始着手,
01:46
and progressing to riskier and harder tasks
26
106708
2185
过渡到风险更高、难度更大的工作,
01:48
as they guide and decide she’s ready.
27
108917
2333
其中确保她准备就绪, 并且有专家在一旁指导。
01:52
My whole life I’ve been fascinated by this kind of learning.
28
112042
2892
我这一生都被这种学习过程所吸引。
01:54
It feels elemental, part of what makes us human.
29
114958
3667
这样基本的步骤,体现了人之常情,
01:59
It has different names: apprenticeship, coaching, mentorship, on the job training.
30
119750
5417
人们为这个过程赋予不同的名字, 学艺、训练、教导和在职培训,
02:05
In surgery, it’s called “see one, do one, teach one.”
31
125542
3291
在外科手术中, 这被称为“看、做、教”,
02:09
But the process is the same,
32
129625
1344
但实际步骤是一样的,
02:10
and it’s been the main path to skill around the globe for thousands of years.
33
130993
4174
这也是千百年来所有人 在培养人才时所用的方式。
02:16
Right now, we’re handling AI in a way that blocks that path.
34
136333
4500
但如今我们应用人工智能的 方法却反其道而行之。
02:21
We’re sacrificing learning in our quest for productivity.
35
141625
2690
为了提高效率,我们 牺牲了学习必经的过程。
02:25
I found this first in surgery while I was at MIT,
36
145292
2809
我在麻省理工学院做手术时 第一次发现了这个现象,
02:28
but now I’ve got evidence it’s happening all over,
37
148125
2476
但现在我发现这样的现象随处可见,
02:30
in very different industries and with very different kinds of AI.
38
150625
3875
遍布各行各业, 以及各项人工智能的应用场景中。
02:35
If we do nothing, millions of us are going to hit a brick wall
39
155083
5851
如果我们无动于衷,成千上万的人 在学习如何掌握人工智能时,
02:40
as we try to learn to deal with AI.
40
160958
2417
将会碰壁。
02:45
Let’s go back to surgery to see how.
41
165125
1772
让我们再用外科手术作为例子,
02:47
Fast forward six months.
42
167708
1935
时间快进六个月,
02:49
It’s 6:30am again, and Kristen is wheeling another prostate patient in,
43
169667
5476
还是凌晨六点半,克里斯汀推着 另一个前列腺病人进手术室。
02:55
but this time to the robotic OR.
44
175167
3166
但这一次,是去自动化手术室。
02:59
The attending leads attaching
45
179667
1684
主治医生把一个
03:01
a four-armed, thousand-pound robot to the patient.
46
181375
2833
长着四只手、重一千镑的 机器人连接到病人身上,
03:04
They both rip off their scrubs,
47
184750
2434
医生们都脱掉了手术服,
03:07
head to control consoles 10 or 15 feet away,
48
187208
3125
来到三五米外的控制台,
03:11
and Kristen just watches.
49
191167
3750
而克里斯汀只负责观察。
03:16
The robot allows the attending to do the whole procedure himself,
50
196375
3053
在机器人的帮助下, 主治医生独自便可完成手术,
03:19
so he basically does.
51
199452
1583
他也是这么做的,
03:21
He knows she needs practice.
52
201917
2101
即使他知道克里斯汀需要练习,
03:24
He wants to give her control.
53
204042
1583
他也希望可以给她机会,
03:26
But he also knows she’d be slower and make more mistakes,
54
206250
3393
但是他同样清楚克里斯汀 操作得更慢,还有失误的风险,
03:29
and his patient comes first.
55
209667
1500
而病人的安全永远是第一位的。
03:32
So Kristin has no hope of getting anywhere near those nerves during this rotation.
56
212250
4625
所以克里斯汀在这次手术中 完全没有机会碰到病人的神经,
03:37
She’ll be lucky if she operates more than 15 minutes during a four-hour procedure.
57
217417
4375
她能在四个小时的手术中 操刀超过一刻钟就算是走运了,
03:42
And she knows that when she slips up,
58
222250
2625
而且她很清楚,万一她出现失误,
03:45
he’ll tap a touch screen, and she’ll be watching again,
59
225458
3042
主治医生就会重新操刀, 她又不得不回到观察者的角色,
03:48
feeling like a kid in the corner with a dunce cap.
60
228917
2625
感到非常沮丧和失落。
03:53
Like all the studies of robots and work I’ve done in the last eight years,
61
233583
3501
正如我过去八年做的 所有关于机器人的研究一样,
在这次研究的开始, 我也提出了一个宏大的问题:
03:57
I started this one with a big, open question:
62
237108
2118
03:59
How do we learn to work with intelligent machines?
63
239250
2792
我们要如何与智能机器共存?
04:02
To find out, I spent two and a half years observing dozens of residents and surgeons
64
242792
5809
为了寻找答案,我花了两年半的时间, 观察了数位外科医生和住院医生。
04:08
doing traditional and robotic surgery, interviewing them
65
248625
3476
他们既做传统的手术, 也做自动化手术,
我采访他们,试图了解他们的学习过程。
04:12
and in general hanging out with the residents as they tried to learn.
66
252125
3338
04:16
I covered 18 of the top US teaching hospitals,
67
256250
3351
这次研究覆盖了 美国18所顶级的教学医院,
04:19
and the story was the same.
68
259625
1458
研究结果显示出相同的趋势。
04:21
Most residents were in Kristen's shoes.
69
261875
2542
大部分住院医生都和克里斯汀一样,
04:24
They got to “see one” plenty,
70
264958
1792
他们“看”得很多,
04:27
but the “do one” was barely available.
71
267583
2292
但“做”的机会却很少。
04:30
So they couldn’t struggle, and they weren’t learning.
72
270333
2528
所以他们难以进步,也无从学习。
04:33
This was important news for surgeons, but I needed to know how widespread it was:
73
273291
3810
这一现象对外科医生来说十分重要, 但我想知道,这样的现象有多普遍?
04:37
Where else was using AI blocking learning on the job?
74
277125
3833
还有哪些领域也是这样, 人工智能阻碍了人们的学习?
04:42
To find out, I’ve connected with a small but growing group of young researchers
75
282208
4310
为了找到答案,我联系了一个 年轻但正迅速成长的研究团队。
04:46
who’ve done boots-on-the-ground studies of work involving AI
76
286542
3434
他们在不同领域都做了一些 关于人工智能的实地研究,
包括初创公司、监管治安部门、
04:50
in very diverse settings like start-ups, policing,
77
290000
2976
投资银行和在线教育等。
04:53
investment banking and online education.
78
293000
2601
04:55
Like me, they spent at least a year and many hundreds of hours observing,
79
295625
5851
和我一样,他们花了至少一年的时间, 用了数百个小时进行观察
05:01
interviewing and often working side-by-side with the people they studied.
80
301500
3917
采访研究对象,甚至 和他们一起生活、工作。
05:06
We shared data, and I looked for patterns.
81
306458
2417
我们共享了数据, 我想从中寻找出规律。
05:09
No matter the industry, the work, the AI, the story was the same.
82
309917
5208
不管在什么行业,利用何种 人工智能,结果都非常相似。
05:16
Organizations were trying harder and harder to get results from AI,
83
316042
3642
企业、机构都卯足了劲, 想从人工智能中获益,
05:19
and they were peeling learners away from expert work as they did it.
84
319708
3542
而这一行为导致学习者 从专业工作中脱离出来。
05:24
Start-up managers were outsourcing their customer contact.
85
324333
2875
初创公司的管理者把 联系消费者的工作外包出去,
05:27
Cops had to learn to deal with crime forecasts without experts support.
86
327833
4042
警察在没有专家的支持下 去做犯罪预测工作,
05:32
Junior bankers were getting cut out of complex analysis,
87
332875
3250
初级银行家被排除在复杂分析之外,
05:36
and professors had to build online courses without help.
88
336500
3083
教授也要独自开始做在线课程。
05:41
And the effect of all of this was the same as in surgery.
89
341125
3226
而这些种种带来的后果 和上述外科例子是一样的,
05:44
Learning on the job was getting much harder.
90
344375
2917
在工作中学习变得越来越难,
05:48
This can’t last.
91
348958
1417
这样的情况需要得到改善。
05:51
McKinsey estimates that between half a billion and a billion of us
92
351542
4267
据麦肯锡估计,到2030年, 我们中有5亿到10亿人,
05:55
are going to have to adapt to AI in our daily work by 2030.
93
355833
4125
将不得不在日常工作中 适应人工智能。
06:01
And we’re assuming that on-the-job learning
94
361000
2011
而我们却以为 在职学习机制将一直存在,
在我们想要学习的时候就唾手可得。
06:03
will be there for us as we try.
95
363035
1917
06:05
Accenture’s latest workers survey showed that most workers learned key skills
96
365500
4268
埃森哲最新的员工调查显示, 多数员工在工作时才能真正掌握技能,
06:09
on the job, not in formal training.
97
369792
2291
而不是在培训中。
06:13
So while we talk a lot about its potential future impact,
98
373292
3517
我们一直在关注 人工智能对未来潜在的影响,
06:16
the aspect of AI that may matter most right now
99
376833
3685
但却忘了它在目前最大的影响,
06:20
is that we’re handling it in a way that blocks learning on the job
100
380542
3375
就是它阻碍了我们学习的步伐,
06:24
just when we need it most.
101
384375
1625
而学习恰恰是 我们目前最需要的东西。
06:27
Now across all our sites, a small minority found a way to learn.
102
387458
6042
现在有一个小群体 找到了学习的方法,
06:35
They did it by breaking and bending rules.
103
395625
3042
通过改变和突破规则。
06:39
Approved methods weren’t working, so they bent and broke rules
104
399083
4643
因为现有的方法不奏效, 所以他们要改变和突破规则,
06:43
to get hands-on practice with experts.
105
403750
1976
来获取和专家一起学习的机会。
06:45
In my setting, residents got involved in robotic surgery in medical school
106
405750
5601
在我经历的环境里,住院医生 在医学院时可以参与到自动化手术中,
06:51
at the expense of their generalist education.
107
411375
3583
牺牲他们的通识教育课程,
06:56
And they spent hundreds of extra hours with simulators and recordings of surgery,
108
416417
5851
他们花了数百个小时 研究模拟器和手术记录,
07:02
when you were supposed to learn in the OR.
109
422292
2541
虽然他们更应该在手术室里实操。
07:05
And maybe most importantly, they found ways to struggle
110
425375
3476
最重要的是, 他们找到了奋斗的方法,
07:08
in live procedures with limited expert supervision.
111
428875
3750
在有限的专家指导下进行现场操作。
07:13
I call all this “shadow learning,” because it bends the rules
112
433792
4309
我称之为“影子学习”, 因为它修改了规则,
07:18
and learner’s do it out of the limelight.
113
438125
2000
让学习者在聚光灯之外学习,
07:21
And everyone turns a blind eye because it gets results.
114
441542
4101
而所有人都对此睁一只眼闭一只眼, 因为这样的学习的确有效。
07:25
Remember, these are the star pupils of the bunch.
115
445667
3166
记住,这样学习的学生都是学霸。
07:29
Now, obviously, this is not OK, and it’s not sustainable.
116
449792
3208
显然,这样的方式并不对, 也并不可持续,
07:33
No one should have to risk getting fired
117
453708
2185
没有人应该要冒着被开除的风险,
07:35
to learn the skills they need to do their job.
118
455917
2150
去学习应掌握的技能,
07:38
But we do need to learn from these people.
119
458792
2056
但我们可能真的要向这些人学习。
07:41
They took serious risks to learn.
120
461917
2250
他们为了学习不惜冒着巨大的风险,
07:44
They understood they needed to protect struggle and challenge in their work
121
464792
4351
他们明白需要保护那些工作中 遇到的困难和挑战,
07:49
so that they could push themselves to tackle hard problems
122
469167
2892
而强迫自己去解决难题,
07:52
right near the edge of their capacity.
123
472083
1959
不断挑战自己的极限。
07:54
They also made sure there was an expert nearby
124
474458
2216
他们也保证身边有 足够的专家资源指导他们,
07:56
to offer pointers and to backstop against catastrophe.
125
476698
3094
在必要的时候出来提供支持。
08:00
Let’s build this combination of struggle and expert support
126
480875
3458
让我们把努力和专家支持结合起来,
08:04
into each AI implementation.
127
484708
2750
将其应用到人工智能中。
08:08
Here’s one clear example I could get of this on the ground.
128
488375
2828
我这里有一个具体的例子,
08:12
Before robots,
129
492125
1226
在有机器人之前,
08:13
if you were a bomb disposal technician, you dealt with an IED by walking up to it.
130
493375
4792
如果你是一个拆弹专家, 你经常要直接处理简单易爆装置,
08:19
A junior officer was hundreds of feet away,
131
499333
2143
一个年轻的警官就在你几百米之外,
08:21
so could only watch and help if you decided it was safe
132
501500
3309
他只能观察你,并且在 你觉得安全的时候才能提供帮助,
08:24
and invited them downrange.
133
504833
1417
才能接近装置。
08:27
Now you sit side-by-side in a bomb-proof truck.
134
507208
3893
现在你们并排坐在防弹卡车里,
08:31
You both watched the video feed.
135
511125
1809
一起看着视频,
08:32
They control a distant robot, and you guide the work out loud.
136
512958
4310
他们远程控制着机器人, 而你大声地指挥工作,
08:37
Trainees learn better than they did before robots.
137
517292
3208
这样一来,他们反而可以 有更好的机会学习。
08:41
We can scale this to surgery, start-ups, policing,
138
521125
3933
我们可以把这种方式应用到 外科手术、初创企业、治安系统、
08:45
investment banking, online education and beyond.
139
525082
2625
投资银行和在线教育等等行业中。
08:48
The good news is we’ve got new tools to do it.
140
528375
2500
好消息是,我们有了 更好的工具辅助学习,
08:51
The internet and the cloud mean we don’t always need one expert for every trainee,
141
531750
4082
网络和云技术的发展意味着我们不再 需要专家进行一对一、面对面的教学,
08:56
for them to be physically near each other or even to be in the same organization.
142
536167
4458
专家和学习者甚至 不需要在同一个机构中。
09:01
And we can build AI to help:
143
541292
3041
我们可以利用人工智能来辅助学习,
09:05
to coach learners as they struggle, to coach experts as they coach
144
545167
5059
在学习者奋斗的过程中指导他们, 还可以指导专家进行更有效的教学,
09:10
and to connect those two groups in smart ways.
145
550250
2542
将两者以更智慧的方式联系起来。
09:15
There are people at work on systems like this,
146
555375
2542
现在已经有在职人员 有这样的教学系统,
09:18
but they’ve been mostly focused on formal training.
147
558333
2792
但是他们也仅仅是 关注入职培训,
09:21
And the deeper crisis is in on-the-job learning.
148
561458
2584
更大的危机其实出现在 在职培训当中。
09:24
We must do better.
149
564417
1851
我们必须要做得更好,
09:26
Today’s problems demand we do better
150
566292
2583
现在出现的问题 要求我们要做得更好,
09:29
to create work that takes full advantage of AI’s amazing capabilities
151
569375
4875
来创造价值,来更好地利用 人工智能带来的便利,
09:35
while enhancing our skills as we do it.
152
575042
2750
同时也让我们的技术变得更加成熟。
09:38
That’s the kind of future I dreamed of as a kid.
153
578333
2750
这才是我小时候梦想的未来,
09:41
And the time to create it is now.
154
581458
2167
而现在正是去开创 这一未来的最佳时机。
09:44
Thank you.
155
584333
1226
谢谢。
09:45
(Applause)
156
585583
3625
(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