听力课堂TED音频栏目主要包括TED演讲的音频MP3及中英双语文稿，供各位英语爱好者学习使用。本文主要内容为演讲MP3+双语文稿:我们如何与人工智能一起工作，希望你会喜欢！

【演讲者及介绍】Matt Beane

是一名技术管理助理教授，在加州大学圣巴巴拉分校的项目与麻省理工学院的数字经济研究所合作。

【演讲主题】我们如何学习与智能机器一起工作

【中英文字幕】

翻译者 Ruijie Wu 校对者 Chen Yunru

00:13

It’s 6:30 in the morning, and Kristen iswheeling her prostate patient into the OR. She's a resident, a surgeon intraining. It’s her job to learn. Today, she’s really hoping to do some of thenerve-sparing, extremely delicate dissection that can preserve erectilefunction. That'll be up to the attending surgeon, though, but he's not thereyet. She and the team put the patient under, and she leads the initialeight-inch incision in the lower abdomen. Once she’s got that clamped back, shetells the nurse to call the attending. He arrives, gowns up, And from there onin, their four hands are mostly in that patient -- with him guiding but Kristinleading the way. When the prostates out (and, yes, he let Kristen do a littlenerve sparing), he rips off his scrubs. He starts to do paperwork. Kristencloses the patient by 8:15, with a junior resident looking over her shoulder.And she lets him do the final line of sutures. Kristen feels great. Patient’sgoing to be fine, and no doubt she’s a better surgeon than she

清晨六点半，克里斯汀正推着她的前列腺病人进手术室。她是一名实习住院外科医生，学习是她工作的一部分。这天，她非常想参与进行神经保留手术，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。不过，这还要看主治医生的意思，但那会儿他并不在手术室。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，当她把切口固定好，便让护士打电话给主治医生。主治医生赶到后，穿上手术服。接着，两人共同开始手术，他们四只手都在病人体内，主治医生负责指导，克里斯汀则主导了手术。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，他脱掉了手术服，开始处理一些文件。而克里斯汀在一个初级住院医生的协助下于8：15完成了手术，克里斯汀还让他给病人做了最后的缝合。克里斯汀感觉好极了，病人很快就会恢复，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. But Kristin’slearning to do her job the way that most of us do: watching an expert for abit, getting involved in easy, safe parts of the work and progressing toriskier and harder tasks as they guide and decide she’s ready. My whole lifeI’ve been fascinated by this kind of learning. It feels elemental, part of whatmakes us human. It has different names: apprenticeship, coaching, mentorship,on the job training. In surgery, it’s called “see one, do one, teach one.” Butthe process is the same, and it’s been the main path to skill around the globefor thousands of years. Right now, we’re handling AI in a way that blocks thatpath. We’re sacrificing learning in our quest for productivity.

虽然，医生的工作挑战性极高。但克里斯汀的学习过程其实和我们的并无分别，通过观察专家的操作，从简单、安全的部分开始着手，过渡到风险更高、难度更大的工作，其中确保她准备就绪，并且有专家在一旁指导。我这一生都被这种学习过程所吸引。这样基本的步骤，体现了人之常情，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，在外科手术中，这被称为“看、做、教”，但实际步骤是一样的，这也是千百年来所有人在培养人才时所用的方式。但如今我们应用人工智能的方法却反其道而行之。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 wasat MIT, but now I’ve got evidence it’s happening all over, in very differentindustries and with very different kinds of AI. If we do nothing, millions ofus are going to hit a brick wall as we try to learn to deal with AI. Let’s goback to surgery to see how.

我在麻省理工学院做手术时第一次发现了这个现象，但现在我发现这样的现象随处可见，遍布各行各业，以及各项人工智能的应用场景中。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，将会碰壁。让我们再用外科手术作为例子，

02:47

Fast forward six months. It’s 6:30am again,and Kristen is wheeling another prostate patient in, but this time to therobotic OR. The attending leads attaching a four-armed, thousand-pound robot tothe patient. They both rip off their scrubs, head to control consoles 10 or 15feet away, and Kristen just watches. The robot allows the attending to do thewhole procedure himself, so he basically does. He knows she needs practice. Hewants to give her control. But he also knows she’d be slower and make moremistakes, and his patient comes first. So Kristin has no hope of gettinganywhere near those nerves during this rotation. She’ll be lucky if sheoperates more than 15 minutes during a four-hour procedure. And she knows thatwhen she slips up, he’ll tap a touch screen, and she’ll be watching again,feeling like a kid in the corner with a dunce cap.

时间快进六个月，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。但这一次，是去自动化手术室。主治医生把一个长着四只手、重一千镑的机器人连接到病人身上，医生们都脱掉了手术服，来到三五米外的控制台，而克里斯汀只负责观察。在机器人的帮助下，主治医生独自便可完成手术，他也是这么做的，即使他知道克里斯汀需要练习，他也希望可以给她机会，但是他同样清楚克里斯汀操作得更慢，还有失误的风险，而病人的安全永远是第一位的。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，她能在四个小时的手术中操刀超过一刻钟就算是走运了，而且她很清楚，万一她出现失误，主治医生就会重新操刀，她又不得不回到观察者的角色，感到非常沮丧和失落。

03:53

Like all the studies of robots and workI’ve done in the last eight years, I started this one with a big, openquestion: How do we learn to work with intelligent machines? To find out, Ispent two and a half years observing dozens of residents and surgeons doingtraditional and robotic surgery, interviewing them and in general hanging outwith the residents as they tried to learn. I covered 18 of the top US teachinghospitals, and the story was the same. Most residents were in Kristen's shoes.They got to “see one” plenty, but the “do one” was barely available. So theycouldn’t struggle, and they weren’t learning.

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：我们要如何与智能机器共存？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。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。这次研究覆盖了美国18所顶级的教学医院，研究结果显示出相同的趋势。大部分住院医生都和克里斯汀一样，他们“看”得很多，但“做”的机会却很少。所以他们难以进步，也无从学习。

04:33

This was important news for surgeons, but Ineeded to know how widespread it was: Where else was using AI blocking learningon the job? To find out, I’ve connected with a small but growing group of youngresearchers who’ve done boots-on-the-ground studies of work involving AI invery diverse settings like start-ups, policing, investment banking and onlineeducation. Like me, they spent at least a year and many hundreds of hoursobserving, interviewing and often working side-by-side with the people theystudied. We shared data, and I looked for patterns. No matter the industry, thework, the AI, the story was the same. Organizations were trying harder andharder to get results from AI, and they were peeling learners away from expertwork as they did it. Start-up managers were outsourcing their customer contact.Cops had to learn to deal with crime forecasts without experts support. Juniorbankers were getting cut out of complex analysis, and professors had to buildonline courses without help. And the effect of all of this was the same as insurgery. Learning on the job was getting much harder.

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？还有哪些领域也是这样，人工智能阻碍了人们的学习？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。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、投资银行和在线教育等。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采访研究对象，甚至和他们一起生活、工作。我们共享了数据，我想从中寻找出规律。不管在什么行业，利用何种人工智能，结果都非常相似。企业、机构都卯足了劲，想从人工智能中获益，而这一行为导致学习者从专业工作中脱离出来。初创公司的管理者把联系消费者的工作外包出去，警察在没有专家的支持下去做犯罪预测工作，初级银行家被排除在复杂分析之外，教授也要独自开始做在线课程。而这些种种带来的后果和上述外科例子是一样的，在工作中学习变得越来越难，

05:48

This can’t last. McKinsey estimates thatbetween half a billion and a billion of us are going to have to adapt to AI inour daily work by 2030. And we’re assuming that on-the-job learning will bethere for us as we try. Accenture’s latest workers survey showed that mostworkers learned key skills on the job, not in formal training. So while we talka lot about its potential future impact, the aspect of AI that may matter mostright now is that we’re handling it in a way that blocks learning on the jobjust when we need it most.

这样的情况需要得到改善。据麦肯锡估计，到2030年，我们中有5亿到10亿人，将不得不在日常工作中适应人工智能。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。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，而不是在培训中。我们一直在关注人工智能对未来潜在的影响，但却忘了它在目前最大的影响，就是它阻碍了我们学习的步伐，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 minorityfound a way to learn. They did it by breaking and bending rules. Approvedmethods weren’t working, so they bent and broke rules to get hands-on practicewith experts. In my setting, residents got involved in robotic surgery inmedical school at the expense of their generalist education. And they spenthundreds of extra hours with simulators and recordings of surgery, when youwere supposed to learn in the OR. And maybe most importantly, they found waysto struggle in live procedures with limited expert supervision. I call all this“shadow learning,” because it bends the rules and learner’s do it out of thelimelight. And everyone turns a blind eye because it gets results. Remember, theseare the star pupils of the bunch.

现在有一个小群体找到了学习的方法，通过改变和突破规则。因为现有的方法不奏效，所以他们要改变和突破规则，来获取和专家一起学习的机会。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，牺牲他们的通识教育课程，他们花了数百个小时研究模拟器和手术记录，虽然他们更应该在手术室里实操。最重要的是，他们找到了奋斗的方法，在有限的专家指导下进行现场操作。我称之为“影子学习”，因为它修改了规则，让学习者在聚光灯之外学习，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。记住，这样学习的学生都是学霸。

07:29

Now, obviously, this is not OK, and it’snot sustainable. No one should have to risk getting fired to learn the skillsthey need to do their job. But we do need to learn from these people. They tookserious risks to learn. They understood they needed to protect struggle andchallenge in their work so that they could push themselves to tackle hardproblems right near the edge of their capacity. They also made sure there wasan expert nearby to offer pointers and to backstop against catastrophe. Let’sbuild this combination of struggle and expert support into each AIimplementation.

显然，这样的方式并不对，也并不可持续，没有人应该要冒着被开除的风险，去学习应掌握的技能，但我们可能真的要向这些人学习。他们为了学习不惜冒着巨大的风险，他们明白需要保护那些工作中遇到的困难和挑战，而强迫自己去解决难题，不断挑战自己的极限。他们也保证身边有足够的专家资源指导他们，在必要的时候出来提供支持。让我们把努力和专家支持结合起来，将其应用到人工智能中。

08:08

Here’s one clear example I could get ofthis on the ground. Before robots, if you were a bomb disposal technician, youdealt with an IED by walking up to it. A junior officer was hundreds of feetaway, so could only watch and help if you decided it was safe and invited themdownrange. Now you sit side-by-side in a bomb-proof truck. You both watched thevideo feed. They control a distant robot, and you guide the work out loud.Trainees learn better than they did before robots. We can scale this tosurgery, start-ups, policing, investment banking, online education and beyond.The good news is we’ve got new tools to do it. The internet and the cloud meanwe don’t always need one expert for every trainee, for them to be physicallynear each other or even to be in the same organization. And we can build AI tohelp: to coach learners as they struggle, to coach experts as they coach and toconnect those two groups in smart ways.

我这里有一个具体的例子，在有机器人之前，如果你是一个拆弹专家，你经常要直接处理简单易爆装置，一个年轻的警官就在你几百米之外，他只能观察你，并且在你觉得安全的时候才能提供帮助，才能接近装置。现在你们并排坐在防弹卡车里，一起看着视频，他们远程控制着机器人，而你大声地指挥工作，这样一来，他们反而可以有更好的机会学习。我们可以把这种方式应用到外科手术、初创企业、治安系统、投资银行和在线教育等等行业中。好消息是，我们有了更好的工具辅助学习，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，专家和学习者甚至不需要在同一个机构中。我们可以利用人工智能来辅助学习，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，将两者以更智慧的方式联系起来。

09:15

There are people at work on systems likethis, but they’ve been mostly focused on formal training. And the deeper crisisis in on-the-job learning. We must do better. Today’s problems demand we dobetter to create work that takes full advantage of AI’s amazing capabilitieswhile enhancing our skills as we do it. That’s the kind of future I dreamed ofas a kid. And the time to create it is now.

现在已经有在职人员有这样的教学系统，但是他们也仅仅是关注入职培训，更大的危机其实出现在在职培训当中。我们必须要做得更好，现在出现的问题要求我们要做得更好，来创造价值，来更好地利用人工智能带来的便利，同时也让我们的技术变得更加成熟。这才是我小时候梦想的未来，而现在正是去开创这一未来的最佳时机。

09:44

Thank you.

谢谢。

09:45

(Applause)

（掌声）