我们的工作将被机器取代，但也有例外The jobs we'll lose to machines -- and the ones we won't | Anthony Goldbloom

So this is my niece.
她叫Yahli
Her name is Yahli.
她只有九个月大
She is nine months old.
她妈妈是一名医生 爸爸是一名律师
Her mum is a doctor, and her dad is a lawyer.
等到Yahli上大学的时候
By the time Yahli goes to college,
像她父母做的这些工作将变得面目全非
the jobs her parents do are going to look dramatically different.
2013年 牛津大学的研究人员 做了一项关于未来就业的研究
In 2013, researchers at Oxford University did a study on the future of work.
他们得出结论：差不多将近一半的工作都有被
They concluded that almost one in every two jobs have a high risk
机器自动化取代的危险
of being automated by machines.
而造成部分人们被取代的原因
Machine learning is the technology
关键就是机器学习技术
that’s responsible for most of this disruption.
它是人工智能最强大的分支
It’s the most powerful branch of artificial intelligence.
它允许机器从现有数据中学习
It allows machines to learn from data
并模仿人类的所作所为
and mimic some of the things that humans can do.
我的公司Kaggle专注于尖端的机器学习
My company, Kaggle, operates on the cutting edge of machine learning.
我们召集了成千上万的专家
We bring together hundreds of thousands of experts
正为工业和学术界寻找重要问题的答案
to solve important problems for industry and academia.
因此 我们可以从独特的视角来观察
This gives us a unique perspective on what machines can do,
机器可以做什么 不可以做什么
what they can’t do
哪些工作可以被自动化或受到威胁
and what jobs they might automate or threaten.
机器学习是在90年代初进入人们的视野
Machine learning started making its way into industry in the early ’90s.
一开始 它只是执行一些相对简单的任务
It started with relatively simple tasks.
像评估贷款申请的信用风险
It started with things like assessing credit risk from loan applications,
通过识别手写的邮政编码来检索邮件
sorting the mail by reading handwritten characters from zip codes.
在过去几年里 我们取得了突破性进展
Over the past few years, we have made dramatic breakthroughs.
现在 机器学习可以完成非常复杂的任务
Machine learning is now capable of far, far more complex tasks.
2012年 Kaggle给当地学校出了个难题
In 2012, Kaggle challenged its community
设计一个算法来评判高中作文
to build an algorithm that could grade high-school essays.
获胜的算法给出的分数居然
The winning algorithms were able to match the grades
和真正老师给出的分数相符
given by human teachers.
去年 我们出了一道更难的题
Last year, we issued an even more difficult challenge.
你能从拍摄出的眼睛图像中 诊断出糖尿病性
Can you take images of the eye and diagnose an eye disease
视网膜病变吗？
called diabetic retinopathy?
再者 获胜的演算法给出的诊断
Again, the winning algorithms were able to match the diagnoses
和眼科医生的诊断相符
given by human ophthalmologists.
类似于这样的任务 只要给定正确的数据
Now, given the right data, machines are going to outperform humans
在这样的任务
at tasks like this.
一位老师在40年的职业生涯中可能审阅一万篇作文
A teacher might read 10,000 essays over a 40-year career.
一名眼科医生 大概可以检查5万只眼睛
An ophthalmologist might see 50,000 eyes.
但在短短几分钟之内 机器可以审阅百万篇文章
A machine can read millions of essays or see millions of eyes
或检查数百万只眼睛
within minutes.
对于频繁、大批量的任务
We have no chance of competing against machines
我们无法与机器抗衡
on frequent, high-volume tasks.
但有些事情机器却无能为力
But there are things we can do that machines can’t do.
机器在解决新情况方面
Where machines have made very little progress
进展甚微
is in tackling novel situations.
它们还不能处理未曾反复接触的事情
They can’t handle things they haven’t seen many times before.
机器学习致命的局限在于
The fundamental limitations of machine learning
它需要从大量已知的数据中总结经验
is that it needs to learn from large volumes of past data.
人类则不然
Now, humans don’t.
我们有一种能把看似毫不相关的事物 联系起来的能力
We have the ability to connect seemingly disparate threads
从而解决从未见过的问题
to solve problems we’ve never seen before.
Percy Spencer是一个物理学家 在二战期间从事雷达的研究工作
Percy Spencer was a physicist working on radar during World War II,
他注意到磁控管融化了他的巧克力
when he noticed the magnetron was melting his chocolate bar.
他从对电磁辐射的理解
He was able to connect his understanding of electromagnetic radiation
联想到烹饪
with his knowledge of cooking
因此发明了——猜猜是什么？微波炉
in order to invent — any guesses? — the microwave oven.
这是个非常杰出的创新例子
Now, this is a particularly remarkable example of creativity.
但这种跨界转型 每天都在以难以察觉的方式成百上千次地
But this sort of cross-pollination happens for each of us in small ways
发生在我们身边
thousands of times per day.
在创新方面
Machines cannot compete with us
机器无法与我们抗衡
when it comes to tackling novel situations,
这将使机器自动化取代
and this puts a fundamental limit on the human tasks
人工的可能性受到限制
that machines will automate.
那么这对未来的工作意味着什么呢？
So what does this mean for the future of work?
未来工作的状态 完全取决于一个问题
The future state of any single job lies in the answer to a single question:
这种工作在多大程度上可以简化为 频繁、大批量的任务
To what extent is that job reducible to frequent, high-volume tasks,
又涉及多少对创新能力的要求？
and to what extent does it involve tackling novel situations?
对于那些频繁 大批量的任务 机器变得越来越智能
On frequent, high-volume tasks, machines are getting smarter and smarter.
如今 它们可以评判作文 诊断某些疾病
Today they grade essays. They diagnose certain diseases.
再过几年 它们将可以进行审计
Over coming years, they’re going to conduct our audits,
将能审阅法律合同样本
and they’re going to read boilerplate from legal contracts.
尽管会计师和律师还是需要的
Accountants and lawyers are still needed.
但他们只需要研究复杂的税收结构
They’re going to be needed for complex tax structuring,
或无先例的诉讼过程
for pathbreaking litigation.
但机器会缩小他们的队伍
But machines will shrink their ranks
并使这些工作机会更难以获得
and make these jobs harder to come by.
如前所述
Now, as mentioned,
在创新方面机器没有取得太大进展
machines are not making progress on novel situations.
营销文案需要抓住消费者的心理
The copy behind a marketing campaign needs to grab consumers’ attention.
它必须从人群中脱颖而出
It has to stand out from the crowd.
经营策略意味着找出那些还没人做的
Business strategy means finding gaps in the market,
市场空白
things that nobody else is doing.
人类将是营销文案的创造者
It will be humans that are creating the copy behind our marketing campaigns,
人类才能推动商业战略发展
and it will be humans that are developing our business strategy.
所以Yahli 无论你将来决定做什么
So Yahli, whatever you decide to do,
让每一天都带给你新的挑战
let every day bring you a new challenge.
如果是那样 你的未来将无法被机器取代
If it does, then you will stay ahead of the machines.
谢谢
Thank you.
[掌声]
[Applause]