人工智能迷局：改变的是什么？
2018-01-26
最近常有一些耸人听闻的新闻标题，诸如：“自动化将使8亿人失业”，“机器人将取代你的工作”，“我们将为机器人工作吗？”
Some of the recent headlines have been dire: “Automation threatens 800 million jobs.” “Robots are coming for jobs.” “Will we work for robots?”.
人工智能和机器学习领域每取得一个进步，人们对于就业前景的预计就愈发悲观。
With every advance in artificial intelligence and machine learning, the predictions of massive displacements in the labor market become more extreme.
但早在几十年前，波兰尼悖论就已经证明，这种忧虑是多余的。
Yet the growing fear of human obsolescence in the workplace is misplaced – as illustrated decades ago in what has come to be known as Polanyi’s Paradox.
1966年，牛津大学著名学者迈克尔·波兰尼（Michael Polanyi）阐明了机器在某些特定领域拥有明显优势，而在另一些领域难以逾越人类。
In 1966, Michael Polanyi, the noted Oxford academic, explained why machines are unbeatable at some tasks and utterly incompetent at others.
他通过评估人类能力，总结道：“我们实际知道的要比我们所能言传的多。”
Assessing human capabilities, he concluded that “we know more than we can tell.”
换句话说，即便人类很擅长做某件事，却无法将具体的做法言之其详。
In other words, humans excel at tasks that we cannot satisfactorily explain how to carry out.
正如我们无法明确表达该如何写诗、如何进行篮球运球，我们也几乎不可能编出对应的程序教会机器怎样做这些事。
The fact that we cannot explain how to do these things – like write a sonnet or dribble a basketball – correlates directly to the near impossibility of programming a machine to do them for us.
基于波兰尼悖论，我们可以作出这样的合理推测：
Because of Polanyi’s Paradox, it is reasonable to expect that,
在很多层面上，人工智能和自动化产业的最新发展将不断重塑劳动力市场的格局，
while the latest wave of advances in AI and automation will reshape the labor market dynamically in many ways,
淘汰某些岗位，并创造出一些现在无法想象的新工种，但它们并不会取代人类。
making certain jobs redundant and creating new ones that we can’t even imagine today, thinking machines are not about to replace human beings.
比如，美联社引进人工智能，运用模板创作了大量的企业报道和赛事简讯，文笔虽不出众，但也在及格线以上。
As an example, at the Associated Press, AI programs now use templates to produce thousands of competent, if not quite elegant, news articles based on corporate earnings or sports scores.
细究一下，我们会发现机器所从事的都是一些记者们避之不及的重复劳动。
Yet when we look closely at what the machines are doing, we see it is mind-numbing, repetitive work that journalists are typically glad to avoid.
新技术给了他们更多的时间外出当面采访 ，更深入地报道时事。
The new technology gives them more time to go out, talk to people and report thoughtfully on what’s happening in the real world.
人工智能领域的重大进展，如IBM Watson以及它在医疗数据分析方面的成功，引发了人们的忧虑，担心人工智能的影响会超过以往的自动化浪潮，也许很快就将取代需要更多技能和培训的工作。
Impressive new developments in AI, like IBM Watson and its success at analyzing medical data, have raised concerns that it might soon be possible to replace workers in jobs that require more skill and training than were affected by past waves of automation.
回到波兰尼悖论，机器现在主要擅长的仍是可以按步骤分解的任务，总有一些工作是人类自己做起来更为得心应手的。
Again we come back to Polanyi’s Paradox. Today, the machines still excel mainly at tasks that can be broken down into a series of discrete steps.
多数主流经济学家认为，自动化的进步将影响未来的工作种类，但不会导致结构性失业。
There will always be room for humans to do the kinds of things that humans do best. Most mainstream economists have concluded that advances in automation will influence the kinds of jobs that are available in the future – but they will not result in structural unemployment.
谈到因技术进步而“消失”的某些工作时，通常只是职位描述已经过时了。
When we hear about jobs that are “lost” as a result of technological change, it is typically only job descriptions that have become obsolete.
“马夫”、“灯夫”这类营生早在20世纪便已绝迹，但他们随即找到了时代中应运而生的新工作。
Occupations like “buggy driver” and “lamp lighter” all but vanished in the 20th century, but workers found employment in new tasks suited to the times.
最近，我81岁的母亲问她的孙子长大后想做什么，他的回答是“也许去开发应用软件”，这让她感到茫然不解。
My 81 year old mother recently asked her grandson what he wishes to become when he grows up. His answer “may be an app developer” left her puzzled.
20世纪80年代出现了电子表格软件，当时人们便普遍预测大批会计将要失业。
When spreadsheet software appeared in the 1980s, it was widely predicted that it would decimate accounting jobs.
而事实上，美国的会计和审计从业人数从1985年的110万上升到2016年的140万。
In fact, the number of accountants and auditors working in the United States rose – from 1.1 million in 1985 to 1.4 million in 2016.
新技术扩展了会计师的工作范围，并对他们的服务提出了新的要求。
The new technology expanded the scope of what accountants could do and generated increased demand for their services.
技术的发展致使人们追求更大的获益，这就是生产力提高所带来的一种重要副产品。
This is one of the critical byproducts of increased productivity – greater demand for the new benefits made possible by technology.
在引入自动化后，对类似煤矿工人这种传统工种的需求量确实会大幅减少。
It is true that some traditional jobs, like in coal mining, now employ a small fraction of the numbers they did before they were automated.
随着就业市场的不断迭代，在当今美国，网站设计师的数量毫无疑问已经超过了煤矿工人。
But the job market keeps evolving. It should come as little surprise that there are now many more people working as web designers in the United States than as coal miners.
如果你对网站设计这类工作不感兴趣，还有很多新的就业机会几乎每天都在涌现——
If web design doesn’t appeal to everyone, other new employment opportunities are opening up almost every day –
比如用户体验设计、纳米技术、可再生能源，甚至还有手工艺品和有机农业。
in fields like user experience design, nanotechnology and renewable energy … or even in artisanal handicrafts and organic farming.
自动化产业本身已经是一个价值2000亿美元的行业，每年的增长率在3%到6%之间。
Automation itself is now a $200 billion industry, growing 3 to 6 percent a year.
可以说，我们公司以及我们的供应商和客户都提供了大量有吸引力的传统职位和新兴职位。
I can attest that, at my company and among our many suppliers and customers, it offers a wide range of compelling traditional and newly emerging jobs.
古往今来，就业市场一直都是动态发展的。
The job market has been historically dynamic and will always keep changing.
作为领导者，我们有责任带领大家与时俱进，在日新月异的世界中为他们提供支持、指引方向。
It is our responsibility as leaders to take the people with us and provide direction and support in the ever faster changing world.
驾驭这些变化的一个关键是与政府、教育界和民营领域的高效协作。
The key to managing these changes is effective cooperation between government, education and the private sector.
我们必须齐心协力，致力于营造一个全新的人类发展生态体系，为有需要的求职者提供培训服务，并倡导所有人终身学习。
We must work together to develop and implement a new ecosystem of people development containing training programs for jobseekers who need them and lifelong development activities for all.
我们要让各级教育能够满足时代的需求，并创造更多的就业机会。
We must ensure that education at all levels keeps up with the needs of employers and the opportunities available to workers.
革新的步伐从未停歇，技术的创新史无前例，要求我们重新思考人类的定位，造福所有人。
The pace of change, the reach and scope of technological innovation is unprecedented and requires a new thinking towards our people to ensure prosperity for everyone.
人们对被自动化抢去生计的恐惧由来已久，但认为人类彻底输给机器的论调是站不住脚的。
The fear of losing one’s livelihood to automation has been with us for a long time. Yet the idea that humans are somehow losing overall to machines is simply inaccurate.
与其杞人忧天担心机器人是否会主宰人类，不如将精力投注于如何掌控机器人、开发人工智能和其它新技术上，以求创造更美好的生活，增加新的就业机会，共创美好未来。
Rather than needless handwringing about the robots becoming our masters, we must channel our energies into mastering robotics, AI and other new technologies for better lives and a new quality of jobs and prosperity.
最近常有一些耸人听闻的新闻标题，诸如：“自动化将使8亿人失业”，“机器人将取代你的工作”，“我们将为机器人工作吗？”
Some of the recent headlines have been dire: “Automation threatens 800 million jobs.” “Robots are coming for jobs.” “Will we work for robots?”.
人工智能和机器学习领域每取得一个进步，人们对于就业前景的预计就愈发悲观。
With every advance in artificial intelligence and machine learning, the predictions of massive displacements in the labor market become more extreme.
但早在几十年前，波兰尼悖论就已经证明，这种忧虑是多余的。
Yet the growing fear of human obsolescence in the workplace is misplaced – as illustrated decades ago in what has come to be known as Polanyi’s Paradox.
1966年，牛津大学著名学者迈克尔·波兰尼（Michael Polanyi）阐明了机器在某些特定领域拥有明显优势，而在另一些领域难以逾越人类。
In 1966, Michael Polanyi, the noted Oxford academic, explained why machines are unbeatable at some tasks and utterly incompetent at others.
他通过评估人类能力，总结道：“我们实际知道的要比我们所能言传的多。”
Assessing human capabilities, he concluded that “we know more than we can tell.”
换句话说，即便人类很擅长做某件事，却无法将具体的做法言之其详。
In other words, humans excel at tasks that we cannot satisfactorily explain how to carry out.
正如我们无法明确表达该如何写诗、如何进行篮球运球，我们也几乎不可能编出对应的程序教会机器怎样做这些事。
The fact that we cannot explain how to do these things – like write a sonnet or dribble a basketball – correlates directly to the near impossibility of programming a machine to do them for us.
基于波兰尼悖论，我们可以作出这样的合理推测：
Because of Polanyi’s Paradox, it is reasonable to expect that,
在很多层面上，人工智能和自动化产业的最新发展将不断重塑劳动力市场的格局，
while the latest wave of advances in AI and automation will reshape the labor market dynamically in many ways,
淘汰某些岗位，并创造出一些现在无法想象的新工种，但它们并不会取代人类。
making certain jobs redundant and creating new ones that we can’t even imagine today, thinking machines are not about to replace human beings.
比如，美联社引进人工智能，运用模板创作了大量的企业报道和赛事简讯，文笔虽不出众，但也在及格线以上。
As an example, at the Associated Press, AI programs now use templates to produce thousands of competent, if not quite elegant, news articles based on corporate earnings or sports scores.
细究一下，我们会发现机器所从事的都是一些记者们避之不及的重复劳动。
Yet when we look closely at what the machines are doing, we see it is mind-numbing, repetitive work that journalists are typically glad to avoid.
新技术给了他们更多的时间外出当面采访 ，更深入地报道时事。
The new technology gives them more time to go out, talk to people and report thoughtfully on what’s happening in the real world.
人工智能领域的重大进展，如IBM Watson以及它在医疗数据分析方面的成功，引发了人们的忧虑，担心人工智能的影响会超过以往的自动化浪潮，也许很快就将取代需要更多技能和培训的工作。
Impressive new developments in AI, like IBM Watson and its success at analyzing medical data, have raised concerns that it might soon be possible to replace workers in jobs that require more skill and training than were affected by past waves of automation.
回到波兰尼悖论，机器现在主要擅长的仍是可以按步骤分解的任务，总有一些工作是人类自己做起来更为得心应手的。
Again we come back to Polanyi’s Paradox. Today, the machines still excel mainly at tasks that can be broken down into a series of discrete steps.
多数主流经济学家认为，自动化的进步将影响未来的工作种类，但不会导致结构性失业。
There will always be room for humans to do the kinds of things that humans do best. Most mainstream economists have concluded that advances in automation will influence the kinds of jobs that are available in the future – but they will not result in structural unemployment.
谈到因技术进步而“消失”的某些工作时，通常只是职位描述已经过时了。
When we hear about jobs that are “lost” as a result of technological change, it is typically only job descriptions that have become obsolete.
“马夫”、“灯夫”这类营生早在20世纪便已绝迹，但他们随即找到了时代中应运而生的新工作。
Occupations like “buggy driver” and “lamp lighter” all but vanished in the 20th century, but workers found employment in new tasks suited to the times.
最近，我81岁的母亲问她的孙子长大后想做什么，他的回答是“也许去开发应用软件”，这让她感到茫然不解。
My 81 year old mother recently asked her grandson what he wishes to become when he grows up. His answer “may be an app developer” left her puzzled.
20世纪80年代出现了电子表格软件，当时人们便普遍预测大批会计将要失业。
When spreadsheet software appeared in the 1980s, it was widely predicted that it would decimate accounting jobs.
而事实上，美国的会计和审计从业人数从1985年的110万上升到2016年的140万。
In fact, the number of accountants and auditors working in the United States rose – from 1.1 million in 1985 to 1.4 million in 2016.
新技术扩展了会计师的工作范围，并对他们的服务提出了新的要求。
The new technology expanded the scope of what accountants could do and generated increased demand for their services.
技术的发展致使人们追求更大的获益，这就是生产力提高所带来的一种重要副产品。
This is one of the critical byproducts of increased productivity – greater demand for the new benefits made possible by technology.
在引入自动化后，对类似煤矿工人这种传统工种的需求量确实会大幅减少。
It is true that some traditional jobs, like in coal mining, now employ a small fraction of the numbers they did before they were automated.
随着就业市场的不断迭代，在当今美国，网站设计师的数量毫无疑问已经超过了煤矿工人。
But the job market keeps evolving. It should come as little surprise that there are now many more people working as web designers in the United States than as coal miners.
如果你对网站设计这类工作不感兴趣，还有很多新的就业机会几乎每天都在涌现——
If web design doesn’t appeal to everyone, other new employment opportunities are opening up almost every day –
比如用户体验设计、纳米技术、可再生能源，甚至还有手工艺品和有机农业。
in fields like user experience design, nanotechnology and renewable energy … or even in artisanal handicrafts and organic farming.
自动化产业本身已经是一个价值2000亿美元的行业，每年的增长率在3%到6%之间。
Automation itself is now a $200 billion industry, growing 3 to 6 percent a year.
可以说，我们公司以及我们的供应商和客户都提供了大量有吸引力的传统职位和新兴职位。
I can attest that, at my company and among our many suppliers and customers, it offers a wide range of compelling traditional and newly emerging jobs.
古往今来，就业市场一直都是动态发展的。
The job market has been historically dynamic and will always keep changing.
作为领导者，我们有责任带领大家与时俱进，在日新月异的世界中为他们提供支持、指引方向。
It is our responsibility as leaders to take the people with us and provide direction and support in the ever faster changing world.
驾驭这些变化的一个关键是与政府、教育界和民营领域的高效协作。
The key to managing these changes is effective cooperation between government, education and the private sector.
我们必须齐心协力，致力于营造一个全新的人类发展生态体系，为有需要的求职者提供培训服务，并倡导所有人终身学习。
We must work together to develop and implement a new ecosystem of people development containing training programs for jobseekers who need them and lifelong development activities for all.
我们要让各级教育能够满足时代的需求，并创造更多的就业机会。
We must ensure that education at all levels keeps up with the needs of employers and the opportunities available to workers.
革新的步伐从未停歇，技术的创新史无前例，要求我们重新思考人类的定位，造福所有人。
The pace of change, the reach and scope of technological innovation is unprecedented and requires a new thinking towards our people to ensure prosperity for everyone.
人们对被自动化抢去生计的恐惧由来已久，但认为人类彻底输给机器的论调是站不住脚的。
The fear of losing one’s livelihood to automation has been with us for a long time. Yet the idea that humans are somehow losing overall to machines is simply inaccurate.
与其杞人忧天担心机器人是否会主宰人类，不如将精力投注于如何掌控机器人、开发人工智能和其它新技术上，以求创造更美好的生活，增加新的就业机会，共创美好未来。
Rather than needless handwringing about the robots becoming our masters, we must channel our energies into mastering robotics, AI and other new technologies for better lives and a new quality of jobs and prosperity.
