How To Build A Human (2016)

Artificial intelligence,
for years a dream of scientists and
Hollywood producers...
Hi. Hello, Caleb.
...but also the violent and
destructive stuff of nightmares.
A brand-new synthetic.
'That's me, Gemma Chan.'
I'm now in set-up mode.
'I play a robot in the futuristic
sci-fi drama, Humans.'
I'm now securely bonded to you
as my primary user.
As the development of artificial
intelligence accelerates and
starts to pervade every
aspect of our lives...
I'm hoping to find out whether the
world depicted in science-fiction
is 10 years away, 100 years away or
closer than we think.
I'm going to meet some of
the greatest minds in science.
They're divided.
Some think AI holds the key to
a safe and prosperous future.
I see AI as the opportunity,
actually,
to unlock humanity's full potential.
Others think the nightmare may
just be about to begin.
We're worried that this will come
out too soon, people will die.
Once you have a kind of
super-intelligent genie that's out
of the bottle, it might not be
possible to put it back in again.
To see just how far we can take
the power of AI, we're conducting
a unique experiment,
building a robot version of me.
This is really the first time we've
tried this, so it's very, very new.
It's really quite uncanny.
'A robot that looks like
me is one thing...' That's my nose.
'..But can it harness the power of
artificial intelligence to
'actually think like me?'
I like the taste of cheese.
'Can we build a human?'
It's so strange.
Hi. Hello. Nice to meet you.
'Day one of the robot build and
we need a body to house its brain.'
Welcome to Millennium FX.
Thank you, thanks for having us.
Millennium FX is one of Europe's
leading suppliers of
prosthetics, animatronics and
specialist make-up.
My goodness. This is something that
we do a lot at Millennium.
You know, babies come up in TV shows
a lot but you can't film on
babies for very long,
so we produce these lifelike babies
so that people can hold them.
So weird!
The closer something is to looking
human, the weirder it feels.
It's what we call
the uncanny valley,
where your mind kind of knows,
kind of knows,
that no matter how perfect
something is,
no matter how good the movement is
or how realistic it looks,
your mind knows that there's
something not quite right.
I think I'm just going to put...put
her down.
'Making a silicon duplicate will
enable the robot builders to
'create a lifelike skin.'
A, E, I, O, U.
'My double will need to have
hundreds of facial expressions just
'like me and they'll need to be in
sync with what the artificial
'brain is thinking.'
Hi, lovely to meet you.
'Time for me to get my face copied.'
We're going to have you 3-D scanned.
Kate Walshe is the producer
in charge.
Today we're going to try and watch
you as closely as possible and see
if we can pick up on all the subtle
expressions and movements of your face.
I'm intrigued to see how far we can
go and how lifelike
we can make it.
I really have
no idea what to expect.
Three, two, one, scanning.
It's so cool.
This is the room where we'll be
doing your head cast and on
the wall behind you are some of
the people who have had
the pleasure of life casting before.
Is that Gordon Ramsay there?
It is, yes.
That's it, perfect.
And if you just want to open
your mouth just
a fraction and just blow out gently.
That's it.
Keep your eyes nice and relaxed.
You're doing fantastically.
The cast consists of two different
types of silicon and is
finished off with
a traditional plaster shell.
One, two, three, great shot.
Well done.
The whole process needs to be quick
as, after 20 minutes, the heat
becomes intolerable.
Just going to whip that off.
That's great. That's a great cast.
You did really, really, really well.
Great, thank you.
'While the team will need to turn
silicon into something that
'resembles me, a bigger challenge
will be to create its mind.
'For that, the robot needs
artificial intelligence,
'defined by computer scientists as
a machine having the capability
'to make a humanlike decision.'
Thank you so much for agreeing
to talk to us today.
'Oxford Professor of philosophy,
Nick Bostrom,
'is one of the world's leading
experts in Al.'
So the goal of AI, artificial
intelligence, has all along,
since its beginnings in the '50s,
been to make machines that have the
same general purpose, smartness that
we humans have, they can do all the
things that the human brain can do.
From smartphones to share prices,
from online customer support
to CCTV, we're
now surrounded by so much of it,
we've started to take it for
granted.
AI is a general-purpose technology.
It can go through any sector -
the economy, it can go through
health care and entertainment,
medicine, defence,
you name it, it could
think of ways in which processes
could be improved by being smarter.
Machines are being developed as
soldiers, mastering real-time
translations using the latest in
word-recognition software.
Can you hear me in French?
TRANSLATED IN FRENCH
And AI could soon have a significant
role in the legal system.
An Al judge that was shadowing
cases from
the European Court of Human Rights
recently came up
with the same decision as
the human judge in four out of five
cases.
The medical profession is also
making use of it.
These Canadian scientists have
recently been to the UK to
sell an AI diagnostic tool that they
say can identify a tumour instantly
and accurately without the need for
a biopsy and with immediate results.
Just like a human, the machine got
trained on multiple thousands
of images to get an accuracy level.
Not only is it less invasive and
quicker but it could save
the NHS millions of pounds
if introduced.
Some of us are already entrusting
our lives to AI.
This is the Tesla Model S on
the market for a cool 100,000.
Noel Sharkey is an academic
specialising in Al and is
interested in how autonomous
decision-making will affect us all.
You're going to take your hands off.
I've arranged to meet him on
a test track to try out one of
the car's most impressive features -
the ability to drive itself
using artificial intelligence.
Right, the hands are coming off.
The hands are off.
Whoa! We're approaching a bend. Oh.
BEEPING
Oh, my goodness!
See, that...it nearly took us off
the road there. Yes!
I know this is meant to be a less
stressful driving experience...
I'm sweating! I'm sweating.
Of course.
I'm sweating. Oh, my goodness.
You try being the passenger!
Semi-autonomous cars like this are
really quite controversial.
Especially in the field of robotics.
And we really need much broader
societal debate.
The UK Government's rolling out
autonomous cars and they
haven't really debated as to whether
people want it or not.
This controversy has been fuelled by
drivers in America ignoring
the company's safety guidelines and
uploading outrageous clips to
the internet.
But I think what puts it all in
perspective is the recent
fatal crash in the USA.
Yes, I remember hearing something
about that in the news.
'Earlier this year,
a driver crashed and died,
'allegedly whilst watching
a Harry Potter film.
'His car hit the underside of
a trailer at 74mph.'
It pulled out right in front of it,
bright white trailer,
bright white sky, very sunny,
and the camera couldn't detect
the trailer and that's why
it ran into it.
And I suppose the more autonomous
vehicles you have on the road,
you know, the likelihood of
incidents like that or
accidents -
they're bound to happen.
It's measuring the space now with
its sensors I think.
But while Tesla say AI cars will
have fewer accidents...
...ultimately, we won't be in
control.
Look,
it's correcting itself as well!
So you imagine what's going to
happen when they really are fully
autonomous because you're going to
have to delegate all your driving
decisions to them.
Life-and-death decisions.
It could well be
life-and-death decisions.
Look, do you see this van
coming down here?
Well, we're just driving along here
on that side in this direction.
The van comes hammering at us for
some reason and you've got
that women there and a pushchair,
and she's going to be hit if
we avoid it.
The car has to make a decision to
take the hit or
to kill the mother and baby.
What would you do?
I don't know! I don't know what
I would do in that split second.
That's a really difficult decision
for anyone to make,
for a human to make. How can a car
decide which life is more valuable?
I don't think it should, myself.
The robot we're building won't need
to make life-and-death decisions.
But it will have to harness Al's
decision-making powers to
think like me.
It's being built on
a modest industrial estate on
the outskirts of Penryn, Cornwall.
I wasn't expecting the hub of
the robot build to be next to a B&Q.
I'm wondering if it's going
to be a...mop head.
'Will Jackson is at the forefront of
constructing humanlike robots.'
Well, hello. Hello.
'To be convincing,
robots need to reason,
'learn and understand so they can
react almost instinctively like us.'
Hello, how are you? I'm very well,
thank you, how are you?
I'm very well, thank you.
He's great.
He's very much a robot,
you definitely know that this is
a machine, not a person.
He's pretty limited.
'Will plans to apply the knowledge
acquired building RoboThespian RT4
'as a basis for
constructing the AI version of me.
'In order to be convincing,
'the robot will need to master
the complexity of language.'
So the first thing that's going to
happen is somebody's going to
speak to the robot and we've got
microphones in the ears that
pick up the sound.
At that point, it's just sound,
it doesn't mean anything.
So we have to turn the sound into
words, into texts.
Once we've got the sound as text,
we've got to try and find
the meaning.
What is this person actually saying
to me? What's the key word?
What did they ask about?
And once we've got that meaning,
we have to try and think of
a sensible reply.
We have to then turn that
back into speech.
So we're going to have to take
a computer-generated version of your
voice that sounds like you.
While we're doing all of this,
the robot can't just sit still.
So it has to have all the little
actions that you would have.
The way you'd listen, the way you
think about what somebody's going
to say, the way you think about what
you're going to say.
You've got to get all these little
subtle things right,
all at the same time.
The brain in this robot has got to
come up with an answer that
makes sense even when it's never
heard the question before.
It's a huge challenge and we have
so many things to get right.
Will's going to use his existing
robot hardware to test out
the speech-recognition software that
he plans to use on the robot me.
There's an old saying in
computer programming,
it's, "Garbage in, garbage out."
If the robot cannot recognise
what you're saying,
if it just gets one or two words in
a sentence wrong when you speak,
what you get back is
complete gibberish.
'Using speech-recognition software,
can it recognise words it
'hears, turn them into text and then
accurately repeat them?'
Echo on.
OK, Echo on.
Peter Piper picked a piece of pickled
pepper, put it on a panda car,
drove it around the moon and ate
a sausage on the way home.
Peter Piper picked a piece of
pickled pepper, put it on a pan,
click and drag random name
generator, sausage on the way home.
"Random name generator sausage"?
Part of the challenge is to get our
robot to respond as quickly as
a human would.
That's within a tenth of a second.
Hello.
Hello. Too slow. Hello.
Hello. It's too slow.
To respond as fast as a human,
the robot needs to work out what's
being said and what it means
before the end of
a sentence and reply with
a lightning-fast response,
which involves instinct like us.
We've got to be so quick with
understanding what's being
said that we can be replying before
even really the last words come out,
so it's got to be superfast.
Building a machine that can
understand a human and answer
back convincingly is one of
the toughest challenges in Al.
If I speak quicker does it
work better?
If I speak quicker does it
work better?
It's almost fast enough, it's almost
there but it's not quite.
If we can't get the recognition that
quick - blown it, I'm robot.
'We're testing the boundaries
of science with
'a unique artificial
intelligence test -
'building a robot that looks,
sounds and thinks like me.'
WHIRRING
The robot team is progressing well,
but it's my facial expressions
that are proving hard for the robot
to sync with its brain.
It's quite a scary looking
thing at the moment,
but once it's got the rest of the
core that goes on which
has got the top teeth in it,
and then the skin will go over
the top, it'll start to look
a lot more like Gemma.
The physical part of the build is
tricky, but it's the robot's ability
to converse like a human that's
really going to test our engineers.
With Moore's Law stating that
computer processing power
doubles every two years,
artificially intelligent
achievements in the real world
have been accelerating.
Fast.
In the 19505, computers beat us
at noughts and crosses.
Then, in the 1990s, they beat
us at chess.
Those computers were programmed to
work out all the possible
outcomes of each move,
and then weighed up how each move
would contribute to
a winning strategy.
But we're now entering an age when
computers aren't just
programmed, but
can learn for themselves.
Computers like IBM's Watson.
This is Jeopardy -
The IBM Challenge.
APPLAUSE
In 2011, Watson was put up to one of
the toughest challenges ever
a general knowledge quiz that
requires logic and quick thinking.
This is Jeopardy. It's
a bit of an American institution.
It's a general knowledge quiz
programme,
and they ask questions in
a really strange way.
They give you the answer and you
have to work out what the
question is.
'Duncan Anderson is IBM's European
chief technology officer for Watson.'
So, we've got the two best
players here.
We've got the person who won the
most amount of money on the show
and the person who has
the longest winning streak.
'Two of the most brilliant brains
had won $5 million between them.
'This game was worth another
million.'
Watson itself is not connected
to the internet,
so it's not out there searching.
It's there, stand-alone, playing
against these champion players.
It was a big risk.
The category is 19th-century
novelists, and here is the clue.
APPLAUSE AND CHEERING
'And this is the point
where Watson's won.'
We've beat the best human
players at Jeopardy.
So, how exactly do you programme
a machine to do something that
it's never done before?
Well, the first thing is you
don't programme it.
Trying to guess every single
question that might come up
and then programme the computer with
the right answer for that question,
we would be here forever.
So, we use this thing called machine
learning, which is an approach
to solving problems whereby the
machine can learn from experiences.
So, we took Watson and we taught it,
we fed it lots of information.
For example, back issues of Time
Magazine, Wikipedia, encyclopaedias.
So, Watson was learning in a way?
Mmm.
And then we go through
a teaching process.
Just like you would teach a child,
we're teaching Watson and we're
testing it, and we're giving it
feedback when it gets it right, and
feedback when it gets it
wrong and then it adjusts its
approach to making decisions.
You could think of it
a bit like trying to find
a pathway through a field.
So, you have a very faint,
distinct path that maybe only one
person has trodden through.
And what you're trying to do is
to feed information
so that that pathway becomes
more defined.
As more people go down that path,
the path gets more trodden through
and becomes more obvious.
So, the more data that you feed into
Watson, it's almost like the...
The more or the wider the
path becomes,
or the more distinct the path
becomes? Exactly.
So, Watson becomes more confident
that that pathway is the
right pathway to take.
'We now need my robot to undergo
a basic version of this process.
'It needs its own pathway, and be
fed hundreds of bespoke new rules
'on how to respond to a question.
'And then learn how to use them.'
Ready to go now you want to try the
latest script from Bruce. OK.
' Key moment. I
will has commissioned one of the
world's leading computer
programmers to build a chat bot - a
piece of software which simulates
human conversation and responds with
the answer it thinks I would give.
Do you like repetition?
Yes. Do you like
repetition?
You said that already. Aha.
Now, that's the kind of reply
I'm looking for.
That's what I want to get to.
That was good. You're improving.
How about we talk about work?
You've played a lot
of supernatural characters.
Are you a fan of the genre or is
it just a coincidence?
I do love science fiction and things
that explore the boundaries
of the possible.
But, actually, having a bunch of
roles in that genre is just
a coincidence.
The team have certainly done
their homework.
Just as Watson was fed information
from previous Jeopardy games, our
chat bot has been fed interviews
and background information on me.
Once we can get the expressions and
all those other little subtle
cues, movement things in there, I
think it'll start to come together.
Fantastic. Yeah,
it will be fantastic.
Are you sure? I am positive.
What's your next question?
'The next question is how to make
it sound like me.'
Hello. Hello. I'm Gemma. And my name
is Bodil. Nice to meet you.
'The human voice has a huge variation
of infection, pitch and intonation.
'Our robot will need to replicate
the essence of my voice,
'with all the specific quirks that
make it unique to me.'
And so just keep in mind that when you
get commas, you make a small pause.
'Bodil Mattison is a computational
linguist and works for
'a company that specialises in
synthesised voices.'
It's not the words that you're
recording here,
you're more interested in the
different sounds that I'm making?
Yeah, I'm not interested in
words at all.
I'm interested in the combinations
of phonemes that we are getting.
Oh, the phonemes? Oh, I see.
So, the combinations of sounds.
How many sentences do I
need to record?
Well, in the end it's about 1,400
sentences. 1,400? Yeah. Today?
This is going to be quite
a long afternoon.
'I certainly have appreciated that
outlet for my creativity.
'Imagine George Bush singing
in the shower.'
You will have to do the
previous one again.
'We're teaching our robot how to
speak like me.'
You are an athlete if you play golf.
'But in the future, we may get to
a stage where it can teach itself -
'learning from experience and coming
up with its own solutions,
'which is what Als are
starting to do.'
Hi. Hi, Demis. Nice to meet you.
Good to meet you. Thanks.
So, welcome to the offices. Thanks
for having me. No problem at all.
'Demis Hassabis was a child
chess prodigy from North London.'
It's actually quite strange meeting
you, cos I've watched you
on screen pretending to be an Al.
Pretending to be a robot? GEMMA LAUGHS
'In 2011, he launched
a British artificial intelligence
'company called Deep Mind.
'Just three years later,
'Google bought his company
for 400 million.'
One of the first things we got our
programmed to do was to play
classic Atari games.
And we wanted the Al to actually
learn to play these games
by just from the pixels on the
screen and no other information,
so it had to learn for itself what
the rules of the game were,
how to get points and how to sort
of master the game.
The idea behind Breakout is that you
control a bat and a ball, and you've
got to break out through a
rainbow-coloured wall brick by brick.
I think I remember this from when I was younger,
yeah. From back when you were playing Atari.
So, you can see it's starting to get
the hang of what it should do.
It's not very good.
It misses the ball quite a lot,
but it's starting to understand that
it's got to move the bat towards
the ball if it wants to get points.
GAME BEEPS
This is after 300 games,
so it's still not that many.
So, we thought this was pretty good,
but what would happen if we
just let the programme continue
to play the game?
So, we left playing for
another 200 games,
and then we came back and it did
this amazing strategy of
digging a tunnel round the side of
the wall and actually sending
the ball round the back.
That's amazing.
It's discovered it for itself,
and obviously can do it, you know,
with superhuman precision.
GAME BEEPS
'Then, last year, Demis and his team
built a computer programme to play
'the most complex game ever devised -
an ancient Chinese game called Go.'
The aim of the game in Go is to
either capture your opponent's
pieces or to surround areas of the
board and make it your territory.
'In chess, the board is made up of
an eight-by-eight grid.
'This means the number of
possible moves in
'a game can be number-crunched by
a computer.
'With a 19-by-19 board,
Go is a much more complex game.'
Even the best players,
they use their intuition and their
instincts more than calculation.
'Astonishingly, the number of
possible moves is greater
'than the number of atoms in
the universe.'
So, even if you took all the world's
computing power and ran it
for a million years, that would
still not be enough to
brute-force a solution to how to win
Go.
'So, Demis gave a more powerful
computer the same challenge
'it gave the Atari computer five
years before.
'Could it teach itself how to play
the world's most complex
'board game and beat the
world's best player?
'Earlier this year,
Demis took the computer programme
'he'd named AlphaGo to Korea to
play the world champion.'
There was actually a genuine, sort
of,
excitement and, sort of, fear about
what was actually going to happen.
'The man on the right is making the
moves on behalf of AlphaGo.'
That's a very... That's
a very surprising move. I thought...
I thought it was... I thought
it was a mistake! HE LAUGHS
AlphaGo played a move that was just
completely unthinkable for
a human to play.
So, there's two important
lines in Go.
If you play on the third line,
you're trying to take territory on
the side of the board.
If you play on the fourth line,
you're trying to take influence into
the centre of the board.
And what of AlphaGo did is it
played on the fifth line.
And you never do that in the
position that it played in.
And we were actually quite worried,
because obviously at that point we
didn't know if this was,
you know, a crazy move or, you know,
a brilliant, original move.
And then 50 moves later,
that move ended up joining up with
another part of the board...
So, it worked? Sort of magically just
resulting in helping it win the game.
'Demis's AI made headlines around
the world when it won the match.'
We're not there yet, but in the
next, you know, few years,
we would like to get to the point
where you could give it any
data, scientific,
medical or commercial,
and it would find the structures or
these patterns that perhaps
human experts have missed,
and highlight those so that
improvements can be made, yeah.
'I think what's really interesting
about it is the fact that
'this programme can teach itself.'
It can learn from its mistakes.
It can come up with a genuinely
creative solution to a problem.
And really you can apply that
to anything.
'So, if we can give my robot the
ability to learn for itself,
'who knows where it might take us?'
It's so strange.
We're testing
the boundaries of science
with a unique
artificial intelligence test -
building a robot that looks,
sounds and thinks like me.
The team is working on getting
the robot's facial expressions
to work in tandem with the AI.
If you look at
other robots of this type,
this kind of flexible silicon-faced
robot, this is bloody good.
Look at me.
Look at me. Oh!
Spooky!
The heart of the robot
is conversation that seems real
and the software to do this,
the chatbot, has just arrived.
Hello. My name is Gemma Chan.
I am not a robot.
That sounds like Gemma, doesn't it?
I think it sounds like Gemma.
Do you think?
I think that sounds like Gemma.
But it's proving a real challenge to
synchronise the body with the mind.
When we can get all of the actions
going at the same time,
it will look really good.
Hopefully the robot will be able
to pick the right expression
for the right thought.
This is the first stored pose
and it's a little smile.
That's awesome.
I like the look of that.
Will is teaching the robot
180 of the most common movements...
I think she's happy.
...hoping she'll choose
the right ones to react
and be convincing as a human.
Gemma now knows
how to smile forever,
and she will be able to seamlessly
blend from one smile to a frown
to anger to despair and
the whole range of human emotions,
but we have to teach her them.
If we can teach AI almost anything that
involves reasoning and decision-making,
once it has those skills,
what might be the consequences?
God, we're high!
This is the London Gateway dockyard.
Every day, over 20,000 containers
are moved by a highly complex AI
which controls
the logistics and timings
of what happens when and where.
I can only see about five people.
Just ten years ago, a port of
this size would have employed
thousands of workers
to shift these containers.
Today, many of the 500 employees
spend their time
supervising the AI machines.
The AI is incredibly efficient,
moving the containers
in the fewest number of moves.
Like a very basic AlphaGo,
it comes up with solutions faster
than any human would be able to.
Wow, that's a lot of containers.
AI expert and writer Martin Ford
thinks what we're seeing here
reflects the shape of
things to come.
Look out at all these containers
here and think of those
as representing the job market
in the United Kingdom
and imagine 35%, roughly
a third of those, disappearing,
and what would happen to our society
and our economy
if that were to happen.
It's an incredible impact on
all of us and on the economy.
It's something that is
going to be uniquely disruptive,
something we've never seen before
in history
and one of the things
that's really driving it is that
machines in a limited way,
are at least, you know,
they're beginning to think.
And it's not just blue-collar jobs.
You may realise this but
a lot of online journalism
based on statistics, like sports
and business articles, are
increasingly being written by Als.
This is a corporate earnings report
for Star Bulk Carriers,
which is actually one of the
companies that utilises this port.
One of these items
is written by a machine
and one is written
by a human journalist.
Just take a look and see if
you can determine which is which.
Yeah, so the first one...
"Athens, Greece. Star Bulk Carriers
Corporation on Wednesday reported
"a loss of $48.8 million
in its first quarter."
Sounds pretty... Yeah,
that sounds pretty human to me.
And the other one starts...
"This Star Bulk Carriers
Corporation reported
"a net revenue decrease of 14.9%
in the first quarter of 2016."
Which one do you think is human and
which one do you think is a machine?
It's really hard to tell.
Which one is written by AI?
Do you know?
I believe the one on the right
is written by a person
and the one on the left
was written by the machine. Really?
They both sound like something
that a person could have written.
That's right. We are really heading
towards a kind of tipping point,
or a point at which things
are going to accelerate
beyond anything we've seen before.
This is just really historic.
It's not just about muscle power
any more. It's about brainpower.
Machines are moving into cognitive
capability and that of course
is the thing that really
sets people apart.
That's the reason that
most people today still have jobs,
whereas horses have been
put out of work.
It's because we have
this ability to think, to learn,
to figure out how to do new things
and to solve problems,
but increasingly the machines
are pushing into that area
and that's going to have huge
implications for the future.
Are any jobs safe?
Right now it's really hard to
build robots that can approach
human ability in dexterity
and mobility,
so a lot of skilled
trade type jobs,
electricians and plumbers
and that type of things,
are probably going to be
relatively safe,
but that's thinking over
the next 10, 20, maybe 30 years.
Once you go beyond that, really,
nothing is off the table.
But the biggest danger
may not be losing our jobs.
Professor Stephen Hawking
recently warned that the creation of
powerful AI will be either
the best or the worst thing
ever to happen to humanity.
Hawking was recently joined by
Tesla founder Elon Musk and other
leading figures in an open letter
highlighting the potential
dangers of unchecked AI.
One of the most vocal
was Professor Nick Bostrom.
Developments in the last few years
in machinery have just been
more rapid than people expected
with these deep learning
algorithms and so forth.
How far away are we from achieving
human level artificial intelligence?
The median opinion - by which year
do you think there's a 50% chance?
There's 2040 or 2050...
Within our lifetime? Yeah.
Within the lifetime of
a lot of people alive today.
The concern there is you are
building this very, very powerful
intelligence and you want
to be really sure then
that this goal that it has
is the same as your goal,
that it incorporates human values
in it,
because if it's not a goal
that you're happy with,
then you might see the world
transform into something that
maximises the Al's goal but leaves
no room for you and your values.
This idea of autonomous and
dangerous Als is a recurring theme
in the world of science fiction.
Are you ever going to let me out?
Yes.
Nick thinks that super intelligent
machines could one day
inhabit the real world and use
their power to negative effect
if we don't put
the right safeguards in place.
Als could take their instructions to
logical but unanticipated extremes.
Ava, I said stop!
The concern is not
that these Als would resent us
or resent being exploited by us
or that they would hate us
or something,
but that they would be
indifferent to us,
so if you think, maybe
you have some big department store
and it wants to build
a new parking place.
Maybe there was an ant colony there
before, right? So it got paved over.
It's not because we hate the ants,
it's just because
they didn't factor into our goal.
I see. And we didn't care.
Similarly, if you had a machine that
wants to optimise the universe
to maximise the realisation
of some goal,
in realising this goal, we wouldn't
be kind of stamped out...
Collateral damage...in the same way
that... Yeah, collateral damage.
The way in which Als can be diverted
from what their architects intended
played out earlier this year when
Microsoft introduced Tay to Twitter.
The AI persona was designed
to act like an American teenager
to attract a younger audience.
The chatbot worked by absorbing
and mimicking the language of
other Twitter users, but
Tay was hijacked by Internet trolls,
who gave it a very different
set of values.
The original intention was corrupted
and Tay was unable to work out
which views were acceptable
and which weren't.
Within a clay, Tay became
a Hitler-loving, racist sex pest.
This shows what can happen to AI
if it falls into the wrong hands
but could a future Tay be far worse?
Once you have a kind of
super-intelligent genie
that's out of the bottle,
it might not be possible to
put it back in again.
You don't want to have a
super-intelligent adversary that is
working at cross purposes with you,
that might then resist
your attempts to shut it down.
It's much better to get it right
on the first attempt, not to
build a super-intelligent evil genie
in the first place, right?
You want to have, if you're going to
have a super-intelligent genie,
you want it to be... You want it to
be on your side. Yeah, exactly.
In Cornwall, our genie is about
to be let out of its bottle,
and I want to know
whose side it's on.
The robot's upstairs here.
Bear in mind this is
not your final skin,
so let's have a look inside.
Let's just let her peep out.
Oh, my goodness.
That's so weird!
See...
It's quite warm.
Just feel it, though. It's weird.
What do you think of the eyes?
Oh. my God!
SHE LAUGHS
It's so strange.
Because she's not quite right,
but she...
You know, I can recognise...
...that the nose is...
Well, I mean, yeah. It's my nose.
Try asking her something.
OK. What have you been up to today?
We've been busy filming
season two of Humans since April.
And it's been very exciting.
Not bad!
So we can have a kind of guess of
the sort of things people might say,
saying...
I can't resist cheese on toast.
What did you have for breakfast?
I had a toasted cheese sandwich.
Is that because
you can't resist cheese on toast?
I like the taste of cheese.
THEY LAUGH
Is that true?
Do you really like cheese on toast?
I love cheese, yeah. Ah!
So, you know,
there's a little personality trait
we might have got right. Maybe not
as much as she likes cheese. Right.
It's the facial expressions.
They're not quite in sync
with what she's saying.
Basically, there's
a slight software bug... Raining.
But don't worry too much...
GEMMA LAUGHS
She's confused! Very strange input.
A few facial tics going on.
Yeah, she has, but, you know,
this is really the first time
we've tried this,
so it's very, very new,
and what you'll find is things
will progress very, very quickly.
Facial tics aside,
the build is going well.
Every day the robot is
making progress in terms of
how it looks, sounds and thinks,
so we've come up with an idea
to put it to the test.
Hey, Gemma. Hi.
Robot Gemma, do not fail.
We're building an artificially
intelligent robot that
looks like me, talks like
me and thinks like me.
And today, I'm going to meet her
in her finished form.
Last time I saw her,
she needed quite a bit of work.
So I'm hoping that today we'll have
more of a finished product.
Yeah, I don't know,
I'm quite nervous.
Oh, no.
Really strange.
It is spooky, isn't it?
SHE GASPS
It's really quite uncanny.
In the 19505, the visionary
godfather of British computing,
Alan Turing,
saw forward to the days when
a computer would be able to think
like us, and he came up with a test.
Could a bystander tell if he was
talking to a machine or a person?
It's known as a Turing test,
and we've adapted it to try out
on our own robot.
This is an enormous challenge.
Basically what we're doing is taking
a fictional vision
of the future and trying to bring
it here now.
She's ready for her close-up.
I've invited some journalists to
come and interview her to see
if they can tell it's not
actually me.
And we've rigged the place with
hidden cameras.
Thank you. Emily.
' Hi there. I
And I've called in a favour from
Humans castmates Emily Berrington
and will Tudor to give proceedings
an air of reality,
and the journalists will
meet them first...
I'll come back and tell you when
we've got two minutes. OK.
...before being taken into
a second room where, to give us
a fighting chance, they'll interview
robot Gemma over Skype.
Hi, Gemma. Hi. I'm
so sorry I can't be there in person.
I'll be in a room down the hall,
watching with the robot's
creators - will and Kate.
Come in. This is very exciting.
'But first, there's just time
to introduce
'robot Gemma to my castmates.'
Oh-ho...
Oh, my goodness, that's terrifying.
What's extraordinary is the tiny
movements in the face.
Are you real? Maybe.
So with everything in place,
the experiment begins.
I'm so nervous. Oh, my God.
So shall I just get started and
ask you questions?
I live in Notting Hill.
That's very interesting.
What do you me...? Could you say...?
Apology accepted.
It's not the smoothest of starts.
My name's Lareb. Got it.
How are you today? Your name is...
Jerome.
No, Lareb, L-A- R- E- B.
Could you say that one more time?
Lareb. Great.
Yeah. What's your name?
Thank you for coming on
and Skype-ing,
that's really nice of you.
Oh, thank you for taking the
time to talk to me.
How was season one for you?
I wish I could explain it to you but
I think it was just an instinct.
Ohh! No...
I think he knows.
So far, robot Gemma's responses
haven't gone as well as I'd hoped.
But gradually, she starts to find
her train of thought.
So, tell me about series two.
It's a key word for Gemma robot,
"series two".
If this journalist just keeps saying
"series two",
everything will go fine.
It seems that the scope
has got bigger for series two.
The main difference is that the
world of the show has become bigger.
And what does Mia want in
series two?
Mia's trying to find her
place in the world.
The voice is a little bit glitchy,
isn't it?
She's going though,
she's getting into it.
Do you think a synth can enjoy art?
That's a good question.
Yes, especially like sculpture.
Yes! Good answer.
Can you hear me now?
Eh, yep, I can see and hear you OK.
Lovely, perfect. OK, so, erm...
We're short on time.
OK, that's absolutely fine.
I don't think she's clocked that
she's not talking to me - yet.
What would you like to see
a robot do?
Oh, I'd love
a robot to tidy my room.
Oh, my God.
What can we expect from season two?
Eh, the synth characters have
fragmented.
In a way. Erm...
Yes, yes...I can't tell you
anything more specific
about the plot I'm afraid.
Can you tell us where we find
your character?
Mia's not with the Hawkins family,
Mia's trying to find her
place in the world.
We've got her.
She totally believes it.
And there's a new character you
spend quite a lot of time with...
She's perfect!
...what can you tell us about their
relationship, if anything?
I can't answer that.
She's not sure. That's OK,
it's OK to be evasive on that.
Who's your favourite character?
You, obviously.
'Time to meet the journalists...
'and reveal the ruse.'
Hello. Hello.
Hi. This is so confusing.
'Just how convincing was it?'
She looks so real! What did you have
for breakfast this morning?
Got dressed in...
Saturday night... Computer.
What do you think? Well, I wasn't
sure, to be honest. Yeah.
What were the giveaways?
The voice, maybe.
It fooled me when I first sat down
and looked at it, yeah,
completely fooled me.
Oh, wow. It just looks like
a kind of not best Skype connection!
THEY LAUGH
How are you today?
I'm very good, thank you.
Did you believe you were
talking to me?
I did, yeah,
I honestly thought it was you.
Could you say that, erm,
one more time?
What's really surprised me is that
we've got as far as we have.
There was obviously something
slightly uncanny about her.
I'm really impressed that
she held her own.
I didn't know if anyone would be
convinced by our robot,
and quite a few people were,
for at least a bit,
so in that sense, that's success.
Having set out to find out how far
we are from
the unsettling
fictional world of Humans,
the answer is perhaps a bit more
complicated than I thought.
Whilst human-like robots may
well be some way away,
what's clear is that Al is
developing at speed,
and we need to debate the potential
pitfalls before it's too late.
It's very clear that these
technologies
are getting better and better.
I do think we have to understand
that we're approaching
this tipping point,
this point where it's going to have
a greatly amplified effect, and we
need to find a way to adapt to that.
The same technology can be used for
good or for bad,
and I think it's down to society and
the inventors of that technology
and the public at large to make sure
it gets used for the right things.
This is just the beginning.