The jobs we'll lose to machines -- and the ones we won't

Machine learning isn't just for simple tasks like assessing credit risk and sorting mail anymore -- today, it's capable of far more complex applications, like grading essays and diagnosing diseases. With these advances comes an uneasy question: Will a robot do your job in the future?

So this is my niece.

Her name is Yahli.

She is nine months old.

Her mum is a doctor,
and her dad is a lawyer.

By the time Yahli goes to college,

the jobs her parents do
are going to look dramatically different.

In 2013, researchers at Oxford University
did a study on the future of work.

They concluded that almost one
in every two jobs have a high risk

of being automated by machines.

Machine learning is the technology

that's responsible for most
of this disruption.

It's the most powerful branch
of artificial intelligence.

It allows machines to learn from data

and mimic some of the things
that humans can do.

My company, Kaggle, operates
on the cutting edge of machine learning.

We bring together
hundreds of thousands of experts

to solve important problems
for industry and academia.

This gives us a unique perspective
on what machines can do,

what they can't do

and what jobs they might
automate or threaten.

Machine learning started making its way
into industry in the early '90s.

It started with relatively simple tasks.

It started with things like assessing
credit risk from loan applications,

sorting the mail by reading
handwritten characters from zip codes.

Over the past few years, we have made
dramatic breakthroughs.

Machine learning is now capable
of far, far more complex tasks.

In 2012, Kaggle challenged its community

to build an algorithm
that could grade high-school essays.

The winning algorithms
were able to match the grades

given by human teachers.

Last year, we issued
an even more difficult challenge.

Can you take images of the eye
and diagnose an eye disease

called diabetic retinopathy?

Again, the winning algorithms
were able to match the diagnoses

given by human ophthalmologists.

Now, given the right data,
machines are going to outperform humans

at tasks like this.

A teacher might read 10,000 essays
over a 40-year career.

An ophthalmologist might see 50,000 eyes.

A machine can read millions of essays
or see millions of eyes

within minutes.

We have no chance of competing
against machines

on frequent, high-volume tasks.

But there are things we can do
that machines can't do.

Where machines have made
very little progress

is in tackling novel situations.

They can't handle things
they haven't seen many times before.

The fundamental limitations
of machine learning

is that it needs to learn
from large volumes of past data.

Now, humans don't.

We have the ability to connect
seemingly disparate threads

to solve problems we've never seen before.

Percy Spencer was a physicist
working on radar during World War II,

when he noticed the magnetron
was melting his chocolate bar.

He was able to connect his understanding
of electromagnetic radiation

with his knowledge of cooking

in order to invent -- any guesses? --
the microwave oven.

Now, this is a particularly remarkable
example of creativity.

But this sort of cross-pollination
happens for each of us in small ways

thousands of times per day.

Machines cannot compete with us

when it comes to tackling
novel situations,

and this puts a fundamental limit
on the human tasks

that machines will automate.

So what does this mean
for the future of work?

The future state of any single job lies
in the answer to a single question:

To what extent is that job reducible
to frequent, high-volume tasks,

and to what extent does it involve
tackling novel situations?

On frequent, high-volume tasks,
machines are getting smarter and smarter.

Today they grade essays.
They diagnose certain diseases.

Over coming years,
they're going to conduct our audits,

and they're going to read boilerplate
from legal contracts.

Accountants and lawyers are still needed.

They're going to be needed
for complex tax structuring,

for pathbreaking litigation.

But machines will shrink their ranks

and make these jobs harder to come by.

Now, as mentioned,

machines are not making progress
on novel situations.

The copy behind a marketing campaign
needs to grab consumers' attention.

It has to stand out from the crowd.

Business strategy means
finding gaps in the market,

things that nobody else is doing.

It will be humans that are creating
the copy behind our marketing campaigns,

and it will be humans that are developing
our business strategy.

So Yahli, whatever you decide to do,

let every day bring you a new challenge.

If it does, then you will stay
ahead of the machines.

Thank you.

(Applause)