Robots & Us: A Brief History of Our Robotic Future

(whirring)

[Narrator] We live in an age of self-driving vehicles,

robots that work alongside us,

and we rely on seemingly omniscient digital systems.

[Computer] Tomorrow in San Francisco it'll be sunny.

[Narrator] Advances in algorithms, sensors

and automation technology stand to upend

nearly every aspect of modern life from work.

Robots are coming out of their cages.

[Narrator] To healthcare.

[Computer] Pigmented lesion benign.

[Narrator] And even how we think of ourselves.

As a tool artificial intelligence might extend human skills

or it might make them obsolete.

Your kitchen knife can be used to make great dinners.

It can be used to kill somebody.

And it's up to you to make that choice.

AI is a tool just like this

except it's a mental tool.

[Narrator] So, how did we get here

and where are we going?

(upbeat music)

As early as the 1940s

the first modern automation began to show up in factories.

Computers were still decades away

from becoming personal

but by the late 50s and early 60s

computer scientists were already trying to figure out

the big question.

Will robots make our lives better

or will they replace us entirely?

On one side was John McCarthy

with artificial intelligence, or AI.

You know, it was roughly the study

of a set of technologies that would sort of mimic

or replicate human capabilities,

whether they were intellectual or physical.

[Narrator] On the other side was Doug Engelbart's idea

of intelligence augmentation, or IA.

And that sort of led to the things

that would become the internet and personal computing.

So, you have these two philosophies

and it's interesting they sort of form

both a dichotomy in a paradox.

If you extend the human in an IA sense of way,

intelligence augmentation, you need fewer humans,

or you can just replace them outright.

[Narrator] Robots and thinking machines

have long been a science fiction fantasy.

But by the 1960s they were actually becoming a reality.

[Narrator] At SRI we are experimenting

with a mobile robot.

We call him Shakey.

[Narrator] A wobbly, incredibly slow reality.

The first real effort to build an autonomous machine

that could move and reason and act in its environment

was Shakey which was a project

proposed by a physicist whose name was Charlie Rosen

at SRI in 1966.

He persuaded the Pentagon

by telling them that they could work on a prototype

of something that might do reconnaissance or be a guard.

At a certain point they asked him

how many guns it might carry and he said

well, two, three, how many do you need?

[Narrator] While Shakey never carried guns

it did mark the beginning

of a new era of computer science.

It was important because it was a platform

on which some of the algorithms

that would later be used by self-driving cars,

the kinds of things that you use in your smartphone,

there was an algorithm called A Star,

that was a navigation algorithm,

is the sort of granddaddy of the way

we get around with our smartphones,

and I think importantly too

that the first work on speech recognition

that was significant was done in Shakey

because they were looking for some way

to interact with the machine.

[Narrator] Since the days of Shakey

computers have advanced remarkably

in their ability to think for themselves.

Advanced AI has beaten humans at their own games,

from chess to Jeopardy.

(clapping)

Perhaps most tangibly of all cars began driving themselves.

12 years ago

the idea that a computer could drive a car

was completely unthinkable.

People felt something as intuitive

and as hard to even explain as driving a car

was reserved for the human race.

[Narrator] Then the Federal Defense Agency DARPA

put on a self-driving car challenge.

Stanford's team, led by Thrun, won

by building Stanley,

a robot car that drove itself 132 miles

across the desert.

Our secret ingredient was AI, it was machine learning.

We actually trained the robot to do the right thing.

Back in the day we trained it how to vary its speed,

where to steer a steering wheel and so on.

[Narrator] Since then AI and advanced automation

have exploded.

Dozens of companies are now testing self-driving vehicles,

most of us carry around smartphones

loaded with AI-powered tools,

and last year Deep Mind's AlphaGo,

a neural network beat us at our hardest game, Go.

That kind of proves to me

that basically everything can be done.

Whatever you say can be done

wait a little and it can be done.

[Narrator] AI and automation promise

faster and safer solutions to humanity's problems.

I see this world where all these basic things

are so affordable and so good to us

that they can free our minds

to develop a humanity that's completely unimaginable today.

[Narrator] But others warn of a jobless future.

It's not just about factories,

and when it is factories,

they're becoming far more advanced,

but it's also, you know, white collar things,

it's jobs dones by journalists and radiologists.

[Narrator] While there have been advances

there are still limitations.

Some of the most cutting-edge autonomous robots

can barely do what a human toddler can do.

This is McCarthy's paradox, John McCarthy,

who was the person who coined the terms AI

noted this originally.

He liked to put his hand into his pocket

and pull out a dime,

and you know, that's something we do

without any thought.

It defies the most sophisticated robotic arm to this day.

[Narrator] But what AI and automation may do

could forever change the world and our place in it.

For the next five episodes of Robots and Us

we'll be exploring how these technologies

could impact everything, from how we work or get around,

to how we take care of our bodies.

[Computer] How did it go with your meds today?

[Narrator] To how we think of ourselves as humans.

(upbeat electronic music)