0:00:02 > 0:00:05There are already nine million robots on our planet.
0:00:05 > 0:00:08You can see something's already happening.
0:00:08 > 0:00:11They are developing so rapidly,
0:00:11 > 0:00:14it's like the arrival of a new species.
0:00:14 > 0:00:16I'm very excited, seeing this.
0:00:20 > 0:00:22What has taken humans millennia...
0:00:22 > 0:00:26robots have achieved in just decades.
0:00:30 > 0:00:34Now they are tackling their greatest challenge -
0:00:34 > 0:00:38trying to think like us.
0:00:38 > 0:00:41I like to learn. This is a ball.
0:00:41 > 0:00:44Brilliant. Oh, he's looking as well.
0:00:44 > 0:00:46I'm Dr Ben Garrod.
0:00:46 > 0:00:48As an evolutionary biologist,
0:00:48 > 0:00:51I'm more used to studying humans and animals.
0:00:52 > 0:00:57So I'm genuinely concerned by how quickly these machines are evolving.
0:00:57 > 0:00:59Konnichiwa.
0:00:59 > 0:01:02Yeah, yeah. Konnichiwa.
0:01:02 > 0:01:06I'm Professor Danielle George. As an electronics engineer,
0:01:06 > 0:01:09I spend a lot of my working life with robots...
0:01:10 > 0:01:12..and I think their rapid development
0:01:12 > 0:01:17provides an incredible opportunity for us all.
0:01:17 > 0:01:21There are literally robots as far as the eye can see,
0:01:21 > 0:01:23and I love it.
0:01:23 > 0:01:26Robots are changing our world.
0:01:27 > 0:01:31In this episode, we investigate whether intelligent robots
0:01:31 > 0:01:34will become our friends and companions...
0:01:34 > 0:01:39She will help set up a home for humans...on Mars.
0:01:39 > 0:01:43..find out if we should trust them with our lives...
0:01:43 > 0:01:45My feet are off, my hands are off!
0:01:45 > 0:01:50..and if one day they will even become conscious.
0:01:50 > 0:01:53Sometimes, the robot does things, we're not sure why it does it.
0:01:55 > 0:01:58Will the rise of robots enhance our lives...
0:01:59 > 0:02:01..or threaten our survival?
0:02:04 > 0:02:05Fire!
0:02:17 > 0:02:19In a laboratory in Southern France,
0:02:19 > 0:02:24we are witnessing yet another robot coming to life.
0:02:24 > 0:02:27- Oh, this thing's incredible. - Wow! Isn't it? Yeah.
0:02:27 > 0:02:30So this is the head. Are you poking the eye right now?
0:02:34 > 0:02:38The robot's design combines artificial intelligence
0:02:38 > 0:02:41with a body based on human anatomy.
0:02:43 > 0:02:46Look at this, this is like... This is like tendons.
0:02:46 > 0:02:48- It's like a proper bone structure. - It really is, the vertebrae
0:02:48 > 0:02:51are all there. You can see the neck, the head.
0:02:51 > 0:02:54I'm guessing the ears are the little red things there.
0:02:54 > 0:02:57This is beautiful, it's aesthetically pleasing, and it's
0:02:57 > 0:03:01been sculpted and crafted in a way that is anatomically beautiful.
0:03:02 > 0:03:05Whilst I am impressed by their design,
0:03:05 > 0:03:09I'm concerned about what the consequences of intelligent robots
0:03:09 > 0:03:11will be for us.
0:03:13 > 0:03:16These things are coming on in leaps and bounds.
0:03:16 > 0:03:17This is designing evolution.
0:03:17 > 0:03:20- Is that freaking you out? - Yeah, suddenly...- Is it?
0:03:20 > 0:03:22Even though it's the same thing, there's a sense of,
0:03:22 > 0:03:25- it's watching me. - And why is that scary?
0:03:25 > 0:03:27It's probably judging me.
0:03:27 > 0:03:28THEY LAUGH
0:03:28 > 0:03:29I don't know! Not as comfortable.
0:03:29 > 0:03:32- I don't want it to look at me any more.- Wow.
0:03:32 > 0:03:35VOICEOVER: I think the development of intelligent robots could help us
0:03:35 > 0:03:39achieve things that we currently find impossible.
0:03:40 > 0:03:41I think it's so funny,
0:03:41 > 0:03:45the way you think that it's going to get this intelligence and do
0:03:45 > 0:03:48the whole Hollywood movie thing of, it's going to take over the world,
0:03:48 > 0:03:50it's going to, like, kill us in our beds, or something.
0:03:50 > 0:03:53No, it's just potentially... Potentially take over the world.
0:03:53 > 0:03:57But the potential, the positive potential these things have
0:03:57 > 0:03:59surely outweighs that?
0:04:00 > 0:04:04Meeting this robot in person has really brought out how unnerved I am
0:04:04 > 0:04:07by the impact of intelligent robots.
0:04:09 > 0:04:12Would it freak you out even more if it talked to you?
0:04:12 > 0:04:15- If that suddenly spoke to me now, I'd be gone!- Really?- Yeah.
0:04:17 > 0:04:21To discover the consequences of robot intelligence on humans,
0:04:21 > 0:04:25we're going to analyse it from a biological perspective.
0:04:28 > 0:04:32We will track down some of the earliest intelligent robots and meet
0:04:32 > 0:04:35more of their modern-day descendants...
0:04:35 > 0:04:38to discover how robot intelligence has evolved
0:04:38 > 0:04:40and where it is really heading.
0:04:48 > 0:04:50My journey begins in the USA...
0:04:51 > 0:04:53..where I'm on my way
0:04:53 > 0:04:57to see one of the most advanced robots in the world.
0:04:58 > 0:05:02I believe intelligent robots WILL become our companions
0:05:02 > 0:05:05and even our friends.
0:05:09 > 0:05:11The hope is that, one day,
0:05:11 > 0:05:14this robot will be working alongside us
0:05:14 > 0:05:17to carry out an extraordinary mission.
0:05:18 > 0:05:21And I can't wait to meet her.
0:05:25 > 0:05:27This is Valkyrie.
0:05:29 > 0:05:32Valkyrie has not been designed for our planet.
0:05:32 > 0:05:37Nasa have created Valkyrie to be an astronaut.
0:05:37 > 0:05:41For a robot, her mission is ambitious.
0:05:41 > 0:05:45In the 2030s, she will help set up a home for humans...
0:05:46 > 0:05:47..on Mars.
0:05:51 > 0:05:54Professor Taskin Padir and his team
0:05:54 > 0:05:58are preparing Valkyrie for her voyage to the red planet.
0:05:58 > 0:06:00So, why is Valkyrie a human form?
0:06:00 > 0:06:03Why isn't it four legs or six legs or one leg or...?
0:06:03 > 0:06:06She will operate in human environments.
0:06:06 > 0:06:09We will design those habitats eventually for astronauts.
0:06:09 > 0:06:12So if she's going to be able to operate in that environment,
0:06:12 > 0:06:14it's better she has the human form factor.
0:06:14 > 0:06:17So, this is a hugely ambitious project, isn't it?
0:06:17 > 0:06:20It is an ambitious project.
0:06:20 > 0:06:24Robots are good for dangerous, distant and daring environments,
0:06:24 > 0:06:26and Mars is a dangerous, distant and daring environment.
0:06:26 > 0:06:31This is really showing the whole thing of humans and robots
0:06:31 > 0:06:33working beside each other,
0:06:33 > 0:06:38and for robots preparing a new environment for humans to live in.
0:06:38 > 0:06:41Yes. Eventually, we really would like to go,
0:06:41 > 0:06:45"All right, Valkyrie, you have landed on an unknown planet.
0:06:45 > 0:06:49"Go figure out the tasks relevant to this space mission."
0:06:49 > 0:06:52I'm in awe of Valkyrie's potential,
0:06:52 > 0:06:55but for her to be a true ally for us in space,
0:06:55 > 0:06:58she needs to be able to operate independently.
0:06:58 > 0:07:02What we're doing today is, line by line,
0:07:02 > 0:07:06we're writing new software so that Valkyrie has better capabilities
0:07:06 > 0:07:09on future missions, specifically to Mars.
0:07:10 > 0:07:13From my background as an electronics engineer,
0:07:13 > 0:07:15I know how difficult this is.
0:07:15 > 0:07:19The team need to write hundreds of thousands of lines of code,
0:07:19 > 0:07:22just for her to carry out the most simple tasks.
0:07:22 > 0:07:26It's time to put Valkyrie to the test.
0:07:26 > 0:07:30We are trying to develop, you know, walking and avoiding obstacles.
0:07:30 > 0:07:35So what she will do is she will walk through and avoid the obstacle,
0:07:35 > 0:07:37then push the button, which will simulate the door opening,
0:07:37 > 0:07:40- and then she'll go through the door.- All right, yeah.
0:07:40 > 0:07:43- So, is she ready to go? - She should be ready to go.
0:07:45 > 0:07:46- OK?- Yeah.- Yeah.
0:07:48 > 0:07:50I'm very excited, seeing this.
0:07:56 > 0:08:00- You can see something's already happening.- Yes, that's the...
0:08:00 > 0:08:04That's the main sensor head on the robot, which maps the environment.
0:08:05 > 0:08:08Valkyrie is equipped with sensors all over her body
0:08:08 > 0:08:10to allow her to navigate.
0:08:14 > 0:08:16- Big girl.- Big girl, yes.
0:08:16 > 0:08:17Yes.
0:08:17 > 0:08:19She takes small steps.
0:08:19 > 0:08:21She's a very clever robot.
0:08:23 > 0:08:25So far, so good.
0:08:25 > 0:08:27Valkyrie has hit the button correctly,
0:08:27 > 0:08:30to simulate opening the capsule door.
0:08:31 > 0:08:34Now she needs to step outside.
0:08:43 > 0:08:45Oh, it was so close!
0:08:48 > 0:08:50Well, she got through the door, yeah.
0:08:50 > 0:08:54She got through the door. That was... That was really good.
0:08:55 > 0:08:57And I guess this is just showing...
0:08:57 > 0:08:59I mean, she... She's at an early stage, isn't she?
0:08:59 > 0:09:03She's at a very early stage. I mean, given that we received the robot
0:09:03 > 0:09:08last year, you know, in one year, we've made quite a bit of progress.
0:09:08 > 0:09:09Yeah. Yeah. She's...
0:09:09 > 0:09:11She's like the baby.
0:09:11 > 0:09:12Yes, that's correct.
0:09:13 > 0:09:16The team try to locate the error...
0:09:18 > 0:09:19Yeah, J1's faulted.
0:09:19 > 0:09:22..and tweak the code so Valkyrie can try again.
0:09:23 > 0:09:26OK, so now the fault's cleared.
0:09:26 > 0:09:28So go ahead and pull J3 and reconfigure.
0:09:28 > 0:09:32- Is the motor power on? - Yeah.- Yeah.
0:09:35 > 0:09:39I'm really struck by the team's enthusiasm and commitment
0:09:39 > 0:09:41to Valkyrie, and I'm interested
0:09:41 > 0:09:44in whether they feel they have a relationship with her.
0:09:46 > 0:09:48Do you become attached to her?
0:09:48 > 0:09:50- Do you have emotion?- We definitely attach personalities.
0:09:50 > 0:09:53You know, we walk in and greet the robot, right?
0:09:53 > 0:09:56- So we say, "Good morning, Valkyrie." - Do you?- And we go from there.
0:09:56 > 0:09:59But Valkyrie has still got a lot to learn.
0:10:04 > 0:10:06- I feel sorry for her. Like, "Oh..." - Exactly, exactly.
0:10:06 > 0:10:08- "She's fallen down!" - And that's why, you know,
0:10:08 > 0:10:10we keep her supported at all times.
0:10:10 > 0:10:13- Yeah.- Because we don't want her to, you know, fall.
0:10:13 > 0:10:16- Like a toddler with reins on, or something.- Yeah.- Makes sense, yeah.
0:10:18 > 0:10:22Anyone who watched Valkyrie take a tumble and thought that was
0:10:22 > 0:10:24some sort of failing of the project,
0:10:24 > 0:10:27to me, has completely missed the point.
0:10:27 > 0:10:30Valkyrie, right now, is like a toddler.
0:10:30 > 0:10:34When my daughter was learning to walk, I expected her to fall.
0:10:34 > 0:10:37For robots like Valkyrie,
0:10:37 > 0:10:39there's bound to be a few tumbles along the way.
0:10:41 > 0:10:46Spending time with Valkyrie confirms my belief that intelligent robots
0:10:46 > 0:10:48WILL help us achieve our dreams.
0:10:50 > 0:10:54But what surprised me was how much the team are already starting
0:10:54 > 0:10:56to bond with her.
0:10:56 > 0:10:58I instantly felt empathy, too.
0:10:58 > 0:11:01I reached out to grab her when she fell.
0:11:05 > 0:11:07But for intelligent robots like Valkyrie
0:11:07 > 0:11:09to truly build relationships,
0:11:09 > 0:11:13they need to engage with us on a much deeper level.
0:11:19 > 0:11:22And that starts with being able to talk to us.
0:11:24 > 0:11:29Talking robots have been the dream of scientists for almost 100 years.
0:11:33 > 0:11:35This is Old Sacramento.
0:11:37 > 0:11:40This place is awesome.
0:11:40 > 0:11:42Look at the buildings here.
0:11:43 > 0:11:48I feel like I've just driven onto the set of a Western movie!
0:11:50 > 0:11:54In this town lives one of the oldest robots in the world.
0:11:55 > 0:11:58His inventor claimed he could respond to the human voice
0:11:58 > 0:11:59and talk back.
0:12:01 > 0:12:02Meet Alpha the robot,
0:12:02 > 0:12:07constructed entirely of metal, but controlled only by the voice.
0:12:07 > 0:12:11I think he could also hold the key to understanding how our robots
0:12:11 > 0:12:12talk to us today.
0:12:14 > 0:12:15How tall are you?
0:12:17 > 0:12:18Six feet.
0:12:18 > 0:12:22Six feet? And how much do you weigh?
0:12:24 > 0:12:26One tonne.
0:12:29 > 0:12:31This is Alpha.
0:12:31 > 0:12:34I really didn't expect to find it
0:12:34 > 0:12:39surrounded here by lots of eclectic memorabilia -
0:12:39 > 0:12:42some bongos in between its legs
0:12:42 > 0:12:46and an animal-skin drinking vessel hanging from its hand.
0:12:46 > 0:12:51This mechanical man was built in Britain in 1932 by an Englishman
0:12:51 > 0:12:53called Harry May.
0:12:54 > 0:12:56Hi, Alpha. How are you today?
0:13:00 > 0:13:01Maybe it's just shy.
0:13:03 > 0:13:07Now he watches over diners in a saloon bar.
0:13:07 > 0:13:10But in his day he was something of a celebrity.
0:13:11 > 0:13:14In 1934, a short film was released
0:13:14 > 0:13:18to showcase Alpha's voice-recognition skills.
0:13:18 > 0:13:21The title was The Face Of Things To Come.
0:13:30 > 0:13:31Not yet.
0:13:35 > 0:13:36Yes.
0:13:36 > 0:13:39People couldn't believe this robot could really talk on its own.
0:13:43 > 0:13:45So, to silence his critics at the time,
0:13:45 > 0:13:49Harry May gave an interview in Time magazine.
0:13:49 > 0:13:53He explained that Alpha's speech was actually just 30 pre-recorded
0:13:53 > 0:13:58responses stored on wax cylinders, like records.
0:13:58 > 0:14:03When Alpha was asked a question, an electronic device inside him
0:14:03 > 0:14:07would decipher the words and select a pre-recorded response to play.
0:14:10 > 0:14:13Harry May's explanation is extraordinary.
0:14:15 > 0:14:17In fact, it's so far ahead of its time
0:14:17 > 0:14:21that I'm intrigued to see exactly what kind of electrical hardware
0:14:21 > 0:14:24is still inside Alpha today.
0:14:26 > 0:14:28Right.
0:14:28 > 0:14:33So, on the archive footage and the photographs that I've seen,
0:14:33 > 0:14:37a lot of the wiring is in the chest, in this chest plate.
0:14:37 > 0:14:38So...
0:14:40 > 0:14:44You can see the bellows and gears.
0:14:44 > 0:14:49I can see some evidence of some great work for the 1930s.
0:14:49 > 0:14:52Unfortunately, I can't find anything
0:14:52 > 0:14:55to substantiate Harry May's explanation.
0:14:57 > 0:15:00So Alpha could have just been controlled by a man hiding behind
0:15:00 > 0:15:05the curtain, operating switches to make his head and mouth move.
0:15:07 > 0:15:11But Harry May had undoubtedly succeeded at tapping into
0:15:11 > 0:15:14our fundamental eagerness to interact with robots.
0:15:16 > 0:15:19And it's that desire to believe in the impossible that drove forward
0:15:19 > 0:15:23the technology of robots for the rest of the 20th century,
0:15:23 > 0:15:27so that today they can talk just like you and me.
0:15:35 > 0:15:38Harry May's vision of humans and robots talking to each other
0:15:38 > 0:15:40was well ahead of its time.
0:15:42 > 0:15:47But his ingenious concept for how Alpha could talk provides one of the
0:15:47 > 0:15:51foundations for how robots communicate with us today.
0:15:52 > 0:15:56I'm driving through the heart of where this technological revolution
0:15:56 > 0:15:57is taking place -
0:15:57 > 0:16:00Silicon Valley.
0:16:00 > 0:16:04It's home to almost every tech start-up you can think of.
0:16:05 > 0:16:10There's a real sense that technology is all around you here,
0:16:10 > 0:16:13a real sense of innovative work going on.
0:16:15 > 0:16:19The tech giants around here have developed their own sophisticated
0:16:19 > 0:16:21intelligent voice assistants.
0:16:21 > 0:16:25And they are designed to be as charming as possible.
0:16:25 > 0:16:28Hey, Siri, do you follow the Three Laws of Robotics?
0:16:29 > 0:16:32Let's see if I can remember.
0:16:32 > 0:16:34OK, I think the three laws are,
0:16:34 > 0:16:36one, clean up your room,
0:16:36 > 0:16:38two, don't run with scissors
0:16:38 > 0:16:40and, three, always wait a half hour
0:16:40 > 0:16:42after eating before going in the water.
0:16:44 > 0:16:47I like Siri. Siri's got a real sense of humour.
0:16:50 > 0:16:53While it might seem like she's talking to me naturally,
0:16:53 > 0:16:57this voice assistant is simply choosing pre-programmed answers
0:16:57 > 0:16:58from its database.
0:17:00 > 0:17:01Do you have a family?
0:17:03 > 0:17:05I have you. That's enough family for me.
0:17:07 > 0:17:11It's just like a massively scaled-up version of the concept Harry May
0:17:11 > 0:17:15described for his robot a century ago,
0:17:15 > 0:17:18where computerised scripts have replaced the wax cylinders.
0:17:18 > 0:17:22But while searching a database for an answer might be useful
0:17:22 > 0:17:24for getting factual information...
0:17:24 > 0:17:29Today, the temperature will range from 11 degrees to 23 degrees.
0:17:29 > 0:17:33..it will take something much more sophisticated for us to have
0:17:33 > 0:17:35a relationship with machines.
0:17:42 > 0:17:45You have to come to Japan to appreciate the deep emotional bonds
0:17:45 > 0:17:48humans can form with robots.
0:17:50 > 0:17:53In Japan, you start to understand what a shared future with robots
0:17:53 > 0:17:57will be like. It's a country that has embraced them like nowhere else,
0:17:57 > 0:17:59a real love affair.
0:18:04 > 0:18:08Here, it's not just important what a robot's voice sounds like,
0:18:08 > 0:18:12but how it is expressed with body language.
0:18:16 > 0:18:19I'm on my way to see a typically Japanese robot.
0:18:21 > 0:18:25It's designed to be our friend and help prevent loneliness.
0:18:35 > 0:18:39In August 2013, a Japanese rocket launched...
0:18:41 > 0:18:44..bound for the International Space Station.
0:18:45 > 0:18:48THEY SHOUT
0:18:49 > 0:18:55Japanese people were so convinced by a robot's ability to be our friend,
0:18:55 > 0:18:59there was a robot astronaut on board.
0:18:59 > 0:19:01His name was Kirobo.
0:19:02 > 0:19:04TRANSLATED:
0:19:06 > 0:19:09Kirobo's mission was to give emotional support
0:19:09 > 0:19:12to Japanese astronaut Koichi Wakata.
0:19:25 > 0:19:27During Kirobo's 18-month stay,
0:19:27 > 0:19:31they shared every experience together,
0:19:31 > 0:19:33even taking selfies.
0:19:51 > 0:19:54I've come to meet Kirobo's creator.
0:19:55 > 0:19:57This is Toyota City.
0:19:57 > 0:19:59It really is a city.
0:19:59 > 0:20:01It's home to over half a million people,
0:20:01 > 0:20:05and roughly 80% of them owe their livelihoods to Toyota.
0:20:08 > 0:20:12I'm here to see why one of the world's largest car-manufacturers
0:20:12 > 0:20:15sent a talking robot into space.
0:20:18 > 0:20:21This is where the original Kirobo was built,
0:20:21 > 0:20:26but today I'm meeting his little brothers and sisters.
0:20:28 > 0:20:30At just four inches tall,
0:20:30 > 0:20:34these guys are the domestic version of their astronaut sibling.
0:20:39 > 0:20:43The developer of the Kirobo family is Hisashi Kusuda.
0:20:44 > 0:20:46Konnichiwa.
0:20:46 > 0:20:48Yeah, yeah. Konnichiwa.
0:20:49 > 0:20:51Konnichiwa.
0:20:51 > 0:20:52Yeah, yeah.
0:20:52 > 0:20:54Konnichiwa.
0:20:54 > 0:20:58These are wonderful robots. Why were they designed?
0:21:04 > 0:21:08Kirobo Minis were designed to tackle the loneliness of modern life
0:21:08 > 0:21:11in a country with an ageing population and falling birth rate.
0:21:13 > 0:21:16Do you think it's possible for a human and a robot,
0:21:16 > 0:21:19like Kirobo, to have a friendship, a bond?
0:21:34 > 0:21:36They do look very human.
0:21:36 > 0:21:39And you designed them like this rather than an animal or an alien
0:21:39 > 0:21:43- or a creature, they are a little person?- Yeah, a person form.
0:21:43 > 0:21:48Kirobo reminds me of babies born in the natural world.
0:21:48 > 0:21:53They often have big heads, large eyes and cute voices
0:21:53 > 0:21:56to help form emotional bonds with their parents.
0:22:03 > 0:22:08Even without understanding, you can see this wonderful communication.
0:22:08 > 0:22:09Traditionally, robots are
0:22:09 > 0:22:13very...clear... in...their...language.
0:22:13 > 0:22:19This one's not, it's quick, it's looking at you, and the intonations,
0:22:19 > 0:22:21it's very humanlike.
0:22:22 > 0:22:27A lot of what I'm seeing also is nonverbal communication,
0:22:27 > 0:22:30so when you and I chat there's a lot of body language.
0:22:35 > 0:22:37- More real.- Yeah.
0:22:39 > 0:22:42Kirobo really has humanlike gestures.
0:22:42 > 0:22:47I feel like I'm interacting with a little person and not a machine.
0:22:49 > 0:22:52But he seems to have one person that he...
0:22:52 > 0:22:55- I don't want to say love, but he has a connection with.- Yes.
0:23:02 > 0:23:03From what I've seen so far,
0:23:03 > 0:23:08this seems to be quite a step in advancing communication with robots,
0:23:08 > 0:23:13this body language, nuances, and it reflects the way we speak,
0:23:13 > 0:23:16the way we communicate. This is helping break down
0:23:16 > 0:23:17a barrier with robots.
0:23:19 > 0:23:22He is, he's a cheeky little icebreaker, he really is.
0:23:28 > 0:23:31Watching Kirobo turn his head and follow my conversation...
0:23:38 > 0:23:40..I can now see much more easily
0:23:40 > 0:23:44a future where we DO have relationships with social robots,
0:23:44 > 0:23:47and that could be helpful to us.
0:23:56 > 0:23:58But despite his clever words,
0:23:58 > 0:24:03Kirobo doesn't really understand me or the world around him.
0:24:03 > 0:24:08If robots really are going to become our friends and, crucially,
0:24:08 > 0:24:12if we are going to trust them,
0:24:12 > 0:24:15they'll have to be able to make sense of our world.
0:24:19 > 0:24:20Like we do.
0:24:25 > 0:24:29And the scale of that problem is staggering,
0:24:29 > 0:24:32even for the most simple task.
0:24:34 > 0:24:36Rush hour in Tokyo.
0:24:36 > 0:24:3938 million people live here.
0:24:41 > 0:24:42And they're on the move.
0:24:43 > 0:24:48How people navigate this megatropolis
0:24:48 > 0:24:51is a wonder of the human brain.
0:24:52 > 0:24:56Shibuya Crossing, the busiest junction in the world.
0:24:56 > 0:24:59Every day, over one million people walk across it.
0:25:02 > 0:25:06A single light change can see some 2,500 commuters battling
0:25:06 > 0:25:10it across, and it lasts just 40 seconds.
0:25:12 > 0:25:16The human brain is the most complex processing machine on the planet.
0:25:16 > 0:25:19And to get me across, mine's receiving data through my eyes,
0:25:19 > 0:25:22my ears and even through my skin,
0:25:22 > 0:25:24and it's using my central nervous system,
0:25:24 > 0:25:25my peripheral nervous system,
0:25:25 > 0:25:29and a brain with over 100 trillion connections.
0:25:30 > 0:25:34But what sets our brains aside from robots and machines is our ability
0:25:34 > 0:25:36to deal with the unpredictable.
0:25:36 > 0:25:37Made it!
0:25:42 > 0:25:44It has taken millions of years
0:25:44 > 0:25:49for the human brain to evolve its beautiful complexity,
0:25:49 > 0:25:52a journey robots have only just begun.
0:25:58 > 0:26:01I've come to Bristol to meet the first robot
0:26:01 > 0:26:04to sense the world around it.
0:26:04 > 0:26:06This was the vital first step needed
0:26:06 > 0:26:10for robots to be able to understand it.
0:26:10 > 0:26:15I'm here to see a man about a tortoise, but not just any tortoise.
0:26:18 > 0:26:23I'm told this little robot was designed with an artificial brain.
0:26:32 > 0:26:36- Hello.- Hi, Ben. Welcome to the lab. - Thank you very much.
0:26:36 > 0:26:39Professor Owen Holland is the world authority on the tortoises.
0:26:41 > 0:26:43Oh, wow!
0:26:44 > 0:26:46Robots everywhere!
0:26:46 > 0:26:47There you are.
0:26:47 > 0:26:50This is brilliant. What have we got here?
0:26:50 > 0:26:51We have Ian, our chief technician,
0:26:51 > 0:26:54who is responsible for the tortoise construction,
0:26:54 > 0:26:56and a couple of tortoises.
0:26:57 > 0:26:59First built in 1948,
0:26:59 > 0:27:03the mastermind behind these artificial animals was neurologist
0:27:03 > 0:27:04Grey Walter.
0:27:06 > 0:27:12His robots had two very basic sensors - sight and touch.
0:27:13 > 0:27:17In a simple villa on the outskirts of Bristol lives Dr Grey Walter,
0:27:17 > 0:27:20who makes robots as a hobby.
0:27:20 > 0:27:22They are small and he doesn't dress them up to look like men.
0:27:22 > 0:27:25He calls them tortoises, and so cunningly have their insides
0:27:25 > 0:27:29been designed that they respond to the stimuli of light and touch
0:27:29 > 0:27:31in a completely lifelike manner.
0:27:33 > 0:27:36This is great to watch. What was he trying to achieve?
0:27:36 > 0:27:40Well, he was a physiologist. He was interested in how the brain worked,
0:27:40 > 0:27:42and he knew that he could never build a model with as many parts
0:27:42 > 0:27:44as the human brain - ten billion -
0:27:44 > 0:27:47so he started thinking maybe it's the number of connections
0:27:47 > 0:27:48that is important.
0:27:48 > 0:27:51This model is named Elsie,
0:27:51 > 0:27:53and she "sees" out of a photoelectric cell
0:27:53 > 0:27:55which rotates above her body.
0:27:55 > 0:27:57When light strikes the cell,
0:27:57 > 0:28:00a driving and steering mechanism sends her hurrying towards it.
0:28:00 > 0:28:03But if she brushes against any object in her path,
0:28:03 > 0:28:06contacts are operated that turn the steering away.
0:28:06 > 0:28:10And so, automatically, she takes avoiding action.
0:28:10 > 0:28:14And so he designed these robots basically to use only two
0:28:14 > 0:28:17what he called nerve cells
0:28:17 > 0:28:20to actually get different types of behaviour,
0:28:20 > 0:28:21and he largely succeeded.
0:28:21 > 0:28:25It's an incredible achievement. Did he make a brain? Was that it?
0:28:25 > 0:28:26The smallest possible brain.
0:28:26 > 0:28:31- Yeah.- But his point was, out of these very few ingredients,
0:28:31 > 0:28:35and some cunning design, you can actually show different behaviours
0:28:35 > 0:28:39- that have the characteristics of life.- So, from your capacity,
0:28:39 > 0:28:42would you say he made synthetic life?
0:28:42 > 0:28:46I would say of a sort. Extremely simple.
0:28:46 > 0:28:48And I regard him as the first pioneer
0:28:48 > 0:28:52of what we call real artificial life. He was building real things
0:28:52 > 0:28:54and saying these are behaving in a lifelike way.
0:28:57 > 0:28:58Looking at his primitive robots
0:28:58 > 0:29:02really does take me back to my student days, when I learned about
0:29:02 > 0:29:05primitive forms of biological life.
0:29:05 > 0:29:08This reminds me of early organisms that had eyes.
0:29:08 > 0:29:10So, things likes snails, trilobites,
0:29:10 > 0:29:12even woodlice that we get in our gardens.
0:29:12 > 0:29:15They're very, very simple eyes, but they're photoreceptive,
0:29:15 > 0:29:17they go towards or away from light,
0:29:17 > 0:29:21- and that's a similar sort of thing here, isn't it?- It is.
0:29:21 > 0:29:24This is like just having one retinal cell.
0:29:24 > 0:29:27It's like having an eye with only one element,
0:29:27 > 0:29:29so it can't do an image or anything like that,
0:29:29 > 0:29:32but it can detect intensity of light, and that's all you need
0:29:32 > 0:29:35to do, that's all the early animals needed to do,
0:29:35 > 0:29:39and this is an example of the robotics following the path.
0:29:39 > 0:29:41Start simple, get complicated.
0:29:41 > 0:29:44It is, and it's responding to the environment. And what you've got
0:29:44 > 0:29:46in evolution that has taken hundreds of millions of years, this is,
0:29:46 > 0:29:48I guess, the first robotic eye,
0:29:48 > 0:29:50and it's evolved from this to where we are now
0:29:50 > 0:29:51within a few decades.
0:29:51 > 0:29:54- Essentially, yes.- It's rapid evolution of robots yet again.
0:29:55 > 0:29:58Owen's robotic tortoises are exact replicas
0:29:58 > 0:30:00of Grey Walter's original design.
0:30:00 > 0:30:03I want to see for myself how they work.
0:30:05 > 0:30:09The touch switch and photoelectric light cell, or robotic eye,
0:30:09 > 0:30:12interact with the circuits controlling the motors,
0:30:12 > 0:30:15enabling the tortoise to drive and turn.
0:30:16 > 0:30:19A proper little retro robot. I'm going to grab this torch...
0:30:19 > 0:30:22- Yeah, we'll see what it can do. - See if he works.- Yeah, OK.
0:30:25 > 0:30:26There we go.
0:30:26 > 0:30:32And what you see is the behaviour when it hasn't detected any light.
0:30:32 > 0:30:34So, it's scanning round all the time.
0:30:34 > 0:30:36We don't like to say it's looking for a light,
0:30:36 > 0:30:38but if it finds a light,
0:30:38 > 0:30:42something will happen. So, if you'd like to switch the torch on...
0:30:42 > 0:30:45Yeah. ..and then point it horizontally...
0:30:45 > 0:30:46Yeah.
0:30:47 > 0:30:50- ..at that.- Oh, wow.
0:30:50 > 0:30:51You will see...
0:30:51 > 0:30:53He's come straight for me.
0:30:53 > 0:30:56And now that it's hit your foot,
0:30:56 > 0:30:58when the touch switch is activated
0:30:58 > 0:31:02it drives forward a bit, turns a bit, forward a bit, turns a bit,
0:31:02 > 0:31:05and this enables it to escape from almost any situation.
0:31:05 > 0:31:09As soon as it's free, it will start scanning for light again.
0:31:09 > 0:31:12This is what gives the impression of intelligence, in that you see
0:31:12 > 0:31:16a sequence of behaviours that, in context, seems to be effective
0:31:16 > 0:31:17and intelligent.
0:31:19 > 0:31:25For its time, this is incredible ingenuity and workmanship, isn't it?
0:31:26 > 0:31:31Grey Walter was one of the first to show that biological principles
0:31:31 > 0:31:35can be applied to the field of robotics.
0:31:36 > 0:31:40Although his double-celled organisms were primitive...
0:31:41 > 0:31:46..they were taking the first steps to make sense of their environment.
0:31:54 > 0:31:58I want to know how far this technology has evolved.
0:31:58 > 0:32:02How close are robots to making sense of the world around them?
0:32:04 > 0:32:06And can we trust their decisions?
0:32:08 > 0:32:13This is quite literally a life-and-death issue for all of us,
0:32:13 > 0:32:17because it's starting to play out on our roads.
0:32:20 > 0:32:23I'm not the most confident road user at the best of times,
0:32:23 > 0:32:26but today I'm having a very different driving experience.
0:32:26 > 0:32:31This may look like a normal vehicle, but it's actually a driverless car,
0:32:31 > 0:32:35and this is a type of robot that's already within our society.
0:32:35 > 0:32:37They're driving on our roads,
0:32:37 > 0:32:39and we're putting our life in their hands,
0:32:39 > 0:32:42so to speak, on a regular basis.
0:32:42 > 0:32:46I've come to Germany and I'm going to let this thing be in control
0:32:46 > 0:32:49as it drives me along one of Germany's busiest roads,
0:32:49 > 0:32:51an autobahn.
0:32:51 > 0:32:53A little bit nervous!
0:32:59 > 0:33:03So this is my first time driving on the left side of a car,
0:33:03 > 0:33:07it's my first time driving, for a long time, an automatic,
0:33:07 > 0:33:10and it's my first time driving in a robot car.
0:33:10 > 0:33:12A day of firsts.
0:33:17 > 0:33:20Joining me on the ride is safety officer Andreas,
0:33:20 > 0:33:24and head of development Dr Miklos Kiss.
0:33:24 > 0:33:26I've got to admit I'm nervous. It's like giving...
0:33:26 > 0:33:28handing over something very precious.
0:33:28 > 0:33:30It's quite a big responsibility, to something,
0:33:30 > 0:33:33and I don't quite know how it works.
0:33:33 > 0:33:36But any second now, I will hand over the controls to Jack,
0:33:36 > 0:33:38my trusty driverless car.
0:33:39 > 0:33:42It's that anticipation. I'm not sure what to expect.
0:33:42 > 0:33:44Right, let's see what happens.
0:33:44 > 0:33:48Come on, Jack. So, I need to press these buttons.
0:33:48 > 0:33:51Keep off your hands from the steering wheel and off your feet...
0:33:51 > 0:33:54- Off the pedals.- My feet are off. My hands are off.- Yeah, that's good.
0:33:54 > 0:33:59- So, Jack is acting. - I love how calm you both are.
0:34:00 > 0:34:04Every instinct in my body has just kicked in,
0:34:04 > 0:34:06and I can actually feel my adrenaline.
0:34:06 > 0:34:09I've gone quite hot and quite sweaty, actually.
0:34:09 > 0:34:13I feel like I'm going to veer off, and I know I won't.
0:34:16 > 0:34:17What I really want to do is...
0:34:17 > 0:34:22- So I can turn around and talk to you now?- Yeah, you can.- And...
0:34:22 > 0:34:25- And that's safe do, obviously, because...- That's safe to do.
0:34:27 > 0:34:30I'm trying very hard not to think about the fact that right now
0:34:30 > 0:34:34my life is in the hands of a robot.
0:34:34 > 0:34:35There's a police car!
0:34:35 > 0:34:37I feel bad - there's a police car in front of me,
0:34:37 > 0:34:39and I haven't got my hands on the wheel!
0:34:39 > 0:34:41Sorry, Officer.
0:34:42 > 0:34:44No-one is controlling this car right now.
0:34:44 > 0:34:47My feet are not controlling any special pedals, my hands are here.
0:34:47 > 0:34:51My eyes are closed, I'm on an autobahn in the middle of Germany.
0:34:51 > 0:34:54It seems so wrong, but I feel so safe.
0:34:54 > 0:34:57And almost like a... Oh, where are we...?
0:34:57 > 0:35:00We indicated! Thanks, Jack, I wasn't concentrating!
0:35:00 > 0:35:04The car's central computer makes sense of the world around it
0:35:04 > 0:35:06using numerous integrated sensors.
0:35:06 > 0:35:08Oh, where are you off to, Jack?
0:35:09 > 0:35:14Those at the front and rear of the car look left and right,
0:35:14 > 0:35:20giving a 360-degree view and a range of 250 metres,
0:35:20 > 0:35:25while a 3-D camera scans traffic conditions and road markings.
0:35:25 > 0:35:29So Jack is constantly sensing every vehicle around us right now,
0:35:29 > 0:35:32I guess in the same way that I'm taking each of my senses
0:35:32 > 0:35:35and getting a holistic view. I guess that's what Jack is doing as well.
0:35:35 > 0:35:37Well, yes.
0:35:38 > 0:35:42The car's computer continuously interprets the data from its sensors
0:35:42 > 0:35:45to generate a 3-D map of the world,
0:35:45 > 0:35:48which it can then safely navigate through.
0:35:48 > 0:35:52It makes split-second decisions to control the braking,
0:35:52 > 0:35:55steering and acceleration.
0:35:55 > 0:35:57That's a huge amount of computational power there.
0:35:57 > 0:36:00What's that comparable to, in terms of other vehicles?
0:36:00 > 0:36:04It's comparable to a military jet.
0:36:04 > 0:36:07So, we're driving something that's comparable to a jet fighter?
0:36:07 > 0:36:09That's it.
0:36:09 > 0:36:11I'm getting over the initial shock
0:36:11 > 0:36:13of actually letting the car take control.
0:36:13 > 0:36:16But I'm still nervous about its judgment.
0:36:16 > 0:36:20I can't quite believe its reactions can be as good as mine.
0:36:20 > 0:36:23So, worst-case scenario, a really, really worst-case scenario,
0:36:23 > 0:36:26somebody turned a car over in front of us now,
0:36:26 > 0:36:28it's 100 metres ahead of us,
0:36:28 > 0:36:30Jack would be able to respond quicker than I could?
0:36:30 > 0:36:32Yeah, quicker than you could.
0:36:32 > 0:36:37So maybe we would be caught in that kind of accident,
0:36:37 > 0:36:40but at least we would do better than a human would.
0:36:40 > 0:36:43- Yeah.- I would like this car to have superhuman power.
0:36:43 > 0:36:46So to solve situations I couldn't do on my own.
0:36:46 > 0:36:49We just had a motorbike go past, we've got vehicles all around us,
0:36:49 > 0:36:52and it's responding easily as well as I could, if...
0:36:52 > 0:36:54as you say, if not better.
0:36:55 > 0:36:58- We've slowed down. That was... - We slowed down.
0:37:01 > 0:37:06I'm really enjoying cruising along this motorway.
0:37:06 > 0:37:08But I've still got some niggling doubts.
0:37:08 > 0:37:13Like, if we did have an accident, who would be responsible?
0:37:14 > 0:37:18This throws up complex ethical and legal questions.
0:37:19 > 0:37:22If we have a crash right now, whose responsibility is it?
0:37:22 > 0:37:24Is it my fault? Is it the car's fault?
0:37:24 > 0:37:26I find it very hard to understand
0:37:26 > 0:37:29- that I wouldn't be responsible if this car crashed.- If the system
0:37:29 > 0:37:31is engaged and accepted it,
0:37:31 > 0:37:35so the handover is done, then the car is responsible.
0:37:35 > 0:37:37So the car means...
0:37:37 > 0:37:40Obviously, if the system does something wrong,
0:37:40 > 0:37:43we at Audi are responsible for what happens.
0:37:44 > 0:37:47There are clearly legal issues to resolve.
0:37:47 > 0:37:49But what's really surprised me
0:37:49 > 0:37:52is that the more I'm being driven around by Jack,
0:37:52 > 0:37:54the more I trust him.
0:37:54 > 0:37:56I'm trusting the car to do its job.
0:37:56 > 0:38:00You are trusting the car to work and to take that responsibility.
0:38:00 > 0:38:03Suddenly, we're putting a lot of trust into...
0:38:03 > 0:38:04- into a robot.- Yeah.
0:38:04 > 0:38:09It's a big step forward, I think, in our social relationship with robots.
0:38:11 > 0:38:15Bizarrely, I do feel comfortable letting a robot take control.
0:38:19 > 0:38:22In a couple of years, we won't think about the robot.
0:38:22 > 0:38:23It will be natural in daily life.
0:38:23 > 0:38:27I think that's the nice part of this.
0:38:27 > 0:38:28My grandmother's in her 90s,
0:38:28 > 0:38:31and she can still remember the first time she saw her very first car.
0:38:31 > 0:38:33And here we are, what, two generations later,
0:38:33 > 0:38:37with me with my hands in the air on an autobahn,
0:38:37 > 0:38:38letting the car drive for me.
0:38:41 > 0:38:45But as much as I have been seduced by the sophistication of the car,
0:38:45 > 0:38:47when we're off the autobahn,
0:38:47 > 0:38:51it also reveals how little Jack and other driverless cars
0:38:51 > 0:38:54truly understand about the world around them.
0:38:54 > 0:38:56Please take over driving.
0:38:56 > 0:38:57So why am I taking over now?
0:38:57 > 0:39:00- What's happening?- Because we are in a construction area,
0:39:00 > 0:39:03and we don't know how the lane markings will be
0:39:03 > 0:39:05and how the side barriers will be.
0:39:05 > 0:39:09- So we don't drive in construction areas right now.- OK.
0:39:10 > 0:39:13Despite all its sensors and computer power,
0:39:13 > 0:39:16without the lane markings of the autobahns,
0:39:16 > 0:39:19Jack can't form an accurate enough 3-D map of the world
0:39:19 > 0:39:22to navigate safely.
0:39:22 > 0:39:24Even I, as a slightly nervous driver,
0:39:24 > 0:39:29still have the ability to understand the world so much better
0:39:29 > 0:39:31than any current driverless car.
0:39:32 > 0:39:35I can not only identify objects,
0:39:35 > 0:39:38I know what things really are and do,
0:39:38 > 0:39:42and that allows me to make profound connections and decisions
0:39:42 > 0:39:46to cope with much more unpredictable scenarios.
0:39:58 > 0:40:01Despite the robot car's limitations,
0:40:01 > 0:40:05I was still amazed to see how far and how fast robots have evolved
0:40:05 > 0:40:09their ability to make sense of the world.
0:40:09 > 0:40:12And I wonder if, one day,
0:40:12 > 0:40:16it will be possible for robots to understand it in the same way we do.
0:40:18 > 0:40:21Can they grasp the true meaning of things
0:40:21 > 0:40:24and develop a sense of self
0:40:24 > 0:40:26to become individuals?
0:40:26 > 0:40:29Yeah? You're going to wave at we now, aren't you?
0:40:31 > 0:40:33Could they even become conscious?
0:40:36 > 0:40:40For humans, the key to our understanding of the world
0:40:40 > 0:40:42is our ability to learn.
0:40:43 > 0:40:46To discover what happens when you try to get a robot to learn
0:40:46 > 0:40:50for itself, I've come to a lab in Japan.
0:40:53 > 0:40:55What have we got going on in here?
0:40:55 > 0:40:57So this is one of our most exciting projects,
0:40:57 > 0:41:01- it's a robot that can learn. - Awesome.
0:41:01 > 0:41:04Can you tell me about the auto focus of this camera?
0:41:09 > 0:41:12This '80s-looking throwback is called Robo V.
0:41:12 > 0:41:13OK...
0:41:13 > 0:41:15For this experiment,
0:41:15 > 0:41:19Professor Dylan Glas has set Robo V a challenge -
0:41:19 > 0:41:23can it learn to be a camera shop salesperson?
0:41:23 > 0:41:27Can you tell me about the auto focus of this camera?
0:41:31 > 0:41:33So, we've got this little robot with one of your colleagues.
0:41:33 > 0:41:36Yeah, so the robot's playing the role of a shopkeeper and it's
0:41:36 > 0:41:39presenting information about the different cameras.
0:41:39 > 0:41:41And the thing we've been exploring lately with this is that the robot
0:41:41 > 0:41:43can actually be proactive.
0:41:43 > 0:41:46So it's not like Siri or something - it's not answering questions.
0:41:46 > 0:41:50It's proactively offering things or suggesting things as well.
0:41:52 > 0:41:54No, I haven't.
0:41:58 > 0:42:00Oh, wow. That's very cool.
0:42:07 > 0:42:10To interact with customers and explain camera functions,
0:42:10 > 0:42:14Robo V is reacting independently.
0:42:14 > 0:42:16Yeah, this does weigh quite heavy.
0:42:18 > 0:42:21What we're exploring here is the concept of,
0:42:21 > 0:42:23how can we program a social robot?
0:42:23 > 0:42:26Instead of classical programming of robots,
0:42:26 > 0:42:28where you program explicitly what the robot should do,
0:42:28 > 0:42:32this robot has learnt everything purely from hundreds of interactions
0:42:32 > 0:42:34that it observed of other people.
0:42:34 > 0:42:37- So this is called learning by imitation.- What's the price?
0:42:41 > 0:42:45Oh, wow. Thank you for your help today.
0:42:48 > 0:42:50To create a Robo V's personality,
0:42:50 > 0:42:54the camera shop scenario was role-played by human shopkeepers
0:42:54 > 0:42:55and customers.
0:42:57 > 0:42:59Hi, this one's 2,000.
0:42:59 > 0:43:03This camera has 18 preset modes.
0:43:03 > 0:43:05Hi, this one is 550.
0:43:05 > 0:43:09- 550? OK, cool. Thank you. - No problem.
0:43:10 > 0:43:14For Robo V to create this database of hundreds of shopkeeper/customer
0:43:14 > 0:43:20interactions, a network of sensors tracked where people moved,
0:43:20 > 0:43:23and microphones captured what they said.
0:43:23 > 0:43:26- This one's 68.- 68?
0:43:26 > 0:43:29- OK, that's really cheap. Thanks. - Yeah, no problem.
0:43:29 > 0:43:31What the robot learns from this is...
0:43:31 > 0:43:33again, this is unsupervised learning,
0:43:33 > 0:43:37it learns on its own to imitate the behaviour that it's shown.
0:43:37 > 0:43:40The locations where people stop in the room, the trajectories
0:43:40 > 0:43:43that people use when they walk to different places,
0:43:43 > 0:43:45it learns all of these things, as well as clusters of speech.
0:43:45 > 0:43:48So maybe you say the same thing in a couple of different ways.
0:43:48 > 0:43:50You might say, "How much is this?"
0:43:50 > 0:43:51"How much is this camera?"
0:43:51 > 0:43:53"How much does this cost?"
0:43:53 > 0:43:55And it will notice that those
0:43:55 > 0:43:58are very similar and cluster them together.
0:43:58 > 0:44:00- ROBO V:- It's got a five-times optical zoom...
0:44:00 > 0:44:02And from this data,
0:44:02 > 0:44:04we had the robot automatically learn
0:44:04 > 0:44:07the logic of how to be the shopkeeper.
0:44:07 > 0:44:11So you've not programmed the robot to be a shopkeeper,
0:44:11 > 0:44:13you've not told it what to say or how to respond,
0:44:13 > 0:44:16it's learned from, effectively, observing the experiences?
0:44:16 > 0:44:19- Exactly.- Can I have a go? - Yeah, please do.
0:44:21 > 0:44:23Can I have some help, please?
0:44:27 > 0:44:29What features does this camera have?
0:44:38 > 0:44:41OK. Can you show me this camera?
0:44:48 > 0:44:50You want me to buy that one, don't you?
0:44:56 > 0:44:58A little bit, yeah.
0:45:02 > 0:45:04I'll come look at that one, then.
0:45:12 > 0:45:15I love this little robot, he's brilliant!
0:45:16 > 0:45:19What's most surprising about my chat with Robo V
0:45:19 > 0:45:23is that this almost feels like an actual conversation
0:45:23 > 0:45:25I would have with a real shopkeeper.
0:45:27 > 0:45:31So this really is a little robot that is behaving just as we would
0:45:31 > 0:45:35in a complex social situation, in a real-world situation.
0:45:35 > 0:45:38So this is, I think, a very powerful concept, because it can scale up.
0:45:38 > 0:45:42If we can capture data of how people interact in the real world
0:45:42 > 0:45:43on a large scale,
0:45:43 > 0:45:47we can use big data to train robots to do very natural interactions.
0:45:47 > 0:45:49Well, instantly, the applications there are massive,
0:45:49 > 0:45:52not only as shopkeepers, but right across the board.
0:45:52 > 0:45:55You've got medical professions or health care, and everything.
0:45:55 > 0:45:57The real challenge is this balance between,
0:45:57 > 0:46:01how controllable is the robot and how much does it learn on its own?
0:46:01 > 0:46:04So sometimes the robot does things, we're not sure why it does it.
0:46:04 > 0:46:05Excuse me?
0:46:05 > 0:46:09ROBOT SPEAKS INDISTINCTLY
0:46:09 > 0:46:12- But, overall, it tends to do pretty good behaviour.- It's fascinating.
0:46:12 > 0:46:15The other interesting thing about this is that the robot doesn't know
0:46:15 > 0:46:19the meaning of anything it does. It's purely behavioural,
0:46:19 > 0:46:22it's purely imitating what it saw the person do before.
0:46:22 > 0:46:25Right, so it's not picking up on keywords -
0:46:25 > 0:46:27"camera" or "cost" or anything like this?
0:46:27 > 0:46:29It doesn't even know anything about English.
0:46:29 > 0:46:34- It's learning through imitation, through experience.- Exactly.- Wow.
0:46:35 > 0:46:40What's blowing my mind is Robo V's behaviour is so humanlike
0:46:40 > 0:46:43that I really believed it had learned to understand
0:46:43 > 0:46:46what I was saying.
0:46:46 > 0:46:49But even if it didn't, does that really matter?
0:46:51 > 0:46:53It can still sell cameras.
0:46:53 > 0:46:57As we move forward, it becomes a philosophically interesting problem,
0:46:57 > 0:47:01because now we're really reflecting on how do we learn? How do we think?
0:47:01 > 0:47:04How do we, you know, ascribe semantic meaning to things,
0:47:04 > 0:47:07and structure, you know, things in the world?
0:47:07 > 0:47:10And these machine learning techniques have provided
0:47:10 > 0:47:12a very interesting lens through which to view
0:47:12 > 0:47:14the way we do our own thoughts.
0:47:14 > 0:47:17So, in the future, I think that these learning systems are really...
0:47:17 > 0:47:21a part of us. Technology is always a part of who we are
0:47:21 > 0:47:25and part of our identity, and this is going to allow us to grow in ways
0:47:25 > 0:47:26we've never been able to grow before.
0:47:30 > 0:47:32It's an extraordinary idea -
0:47:32 > 0:47:36that in trying to teach robots to learn human cognitive abilities,
0:47:36 > 0:47:41we may also learn more about how we think ourselves.
0:47:44 > 0:47:46The key to this may be to teach robots
0:47:46 > 0:47:49not to simply mimic our behaviour
0:47:49 > 0:47:52but to develop a conceptual understanding of the world
0:47:52 > 0:47:54for themselves,
0:47:54 > 0:47:59so they can generate humanlike thought and behaviour spontaneously.
0:48:03 > 0:48:05I've come to Plymouth University's
0:48:05 > 0:48:07Centre for Robotics and Neural Systems
0:48:07 > 0:48:11to meet a team of scientists that is trying to do just that.
0:48:13 > 0:48:16Their robot is called iCub.
0:48:16 > 0:48:18- There he is.- This is the iCub.
0:48:18 > 0:48:21The famous little iCub.
0:48:21 > 0:48:23And I say little, I mean, it's astounding just how much
0:48:23 > 0:48:26he resembles - I keep saying "he" already - a small child.
0:48:29 > 0:48:33At one metre tall and weighing 22 kilos,
0:48:33 > 0:48:38iCub not only looks like a child, but learns like one, too.
0:48:39 > 0:48:43Angelo Cangelosi, Professor of Artificial Intelligence
0:48:43 > 0:48:45and Cognition, is his guardian.
0:48:45 > 0:48:48It's almost like a two-year-old child, and in fact,
0:48:48 > 0:48:50like a two-year-old child,
0:48:50 > 0:48:53we're going to teach it the name of objects, one word at a time.
0:48:53 > 0:48:57That's what children do between 1½ years of age and two years.
0:48:57 > 0:49:00You say "teach" - how does it learn? What has it got in there?
0:49:00 > 0:49:03The robot has a simulated brain and, as the brain of a child,
0:49:03 > 0:49:06is able to associate, to learn the correspondences
0:49:06 > 0:49:10between the sound of a word and the picture of an object.
0:49:12 > 0:49:15iCub is equipped with cameras to see...
0:49:17 > 0:49:19..microphones to hear...
0:49:19 > 0:49:21and even smart skin to touch.
0:49:24 > 0:49:27The information it gathers from the stimuli around it
0:49:27 > 0:49:30is fed into an artificial neural network -
0:49:30 > 0:49:34a computer system inspired by the human brain.
0:49:37 > 0:49:40iCub is not simply mimicking human behaviour...
0:49:42 > 0:49:44..it is trying to discover for itself
0:49:44 > 0:49:48the relationships between what it can see,
0:49:48 > 0:49:51what it can hear and what it can touch,
0:49:51 > 0:49:52just like a child.
0:49:53 > 0:49:55I want to see how it learns.
0:49:57 > 0:50:01Right. OK, iCub, let's put the ball there so you can see.
0:50:02 > 0:50:04Learn "ball".
0:50:05 > 0:50:07I like to learn.
0:50:07 > 0:50:08This is a ball.
0:50:08 > 0:50:11OK. Brilliant.
0:50:11 > 0:50:12Let's try...
0:50:13 > 0:50:15Try another one. What have we got here?
0:50:17 > 0:50:19OK, iCub.
0:50:19 > 0:50:20Learn "cup".
0:50:21 > 0:50:23I like to learn.
0:50:23 > 0:50:24This is a cup.
0:50:24 > 0:50:26Well done. Right.
0:50:26 > 0:50:28OK, so we've taught iCub two new objects.
0:50:28 > 0:50:31How do I know if he's actually learned this or not?
0:50:31 > 0:50:32Let's ask him to name them.
0:50:32 > 0:50:36- Right, OK.- So if you show an object, you can then ask for the name.
0:50:38 > 0:50:39What's this?
0:50:41 > 0:50:43It should be a cup.
0:50:43 > 0:50:45It is a cup! Well done.
0:50:45 > 0:50:48OK. I'm going to really test him.
0:50:48 > 0:50:50and see if he can find the one I'm asking for.
0:50:52 > 0:50:53That's close enough.
0:50:55 > 0:50:58OK, iCub, find cup.
0:50:59 > 0:51:03OK. Now I'm looking for a cup.
0:51:03 > 0:51:06Oh, his eyes are moving, his head's moving...
0:51:06 > 0:51:09- Yes.- ..and he's tracking the cup. He's not interested in the ball.
0:51:09 > 0:51:11Try to show the ball also.
0:51:13 > 0:51:16He doesn't care. He wants...
0:51:16 > 0:51:19- He likes the cup.- I mean, his eyes... He wants the cup.
0:51:19 > 0:51:21He's not interested in the ball whatsoever.
0:51:22 > 0:51:25This is because he had learned the two objects,
0:51:25 > 0:51:30and therefore it's following what we ask it to do.
0:51:30 > 0:51:33This is incredible. So we've literally
0:51:33 > 0:51:36just taught this cute little robot...
0:51:36 > 0:51:40..A two-year-old robot the names of objects, like a two-year-old child.
0:51:42 > 0:51:45As toddlers interact with the world around them,
0:51:45 > 0:51:48they learn from one experience to the next,
0:51:48 > 0:51:51making connections between what they can see and hear
0:51:51 > 0:51:54to form the basis of context and meaning.
0:51:54 > 0:51:59These become the building blocks of intelligence and reasoning.
0:52:00 > 0:52:02There are things which are harder.
0:52:02 > 0:52:06You can recognise a cup because of its shape and its colour,
0:52:06 > 0:52:07you can recognise a ball, again,
0:52:07 > 0:52:10because of the different shape compared to a cup.
0:52:10 > 0:52:12But what about teaching a robot or a child
0:52:12 > 0:52:15to understand the number one and number two.
0:52:15 > 0:52:17How would you do this?
0:52:18 > 0:52:19One, two, three?
0:52:20 > 0:52:23And, like us, the more we learn,
0:52:23 > 0:52:25the more complex the tasks we can tackle.
0:52:27 > 0:52:28Oh, he's looking as well.
0:52:31 > 0:52:32One.
0:52:36 > 0:52:38Two.
0:52:38 > 0:52:41This is great. What's going on behind the scenes as he's counting?
0:52:41 > 0:52:43- Three.- We have a brain, an artificial brain,
0:52:43 > 0:52:46that's been trained to learn to associate sounds,
0:52:46 > 0:52:49number words in this case, with its finger position.
0:52:50 > 0:52:54By doing this, the robot is actually able to use its body
0:52:54 > 0:52:56to learn that there are sequences which are fixed.
0:52:56 > 0:53:02For example, one comes before two, two before three, and so on.
0:53:02 > 0:53:04I guess that's why iCub is so special,
0:53:04 > 0:53:06because you've got that wonderful integration
0:53:06 > 0:53:09between the cognitive capability up in here,
0:53:09 > 0:53:11but also that physical embodiment.
0:53:11 > 0:53:13You've got the two things combined, haven't you?
0:53:13 > 0:53:16This really shows why a body is important for a robot,
0:53:16 > 0:53:19the same way a body is important for a child.
0:53:19 > 0:53:22Children learn by using their motor skills to explore
0:53:22 > 0:53:26the physical world around them through touch and movement.
0:53:26 > 0:53:29As their body interacts with the environment,
0:53:29 > 0:53:32they learn from each new experience.
0:53:33 > 0:53:35iCub does the same.
0:53:35 > 0:53:37In tiny little steps,
0:53:37 > 0:53:41it is trying to form its own unique understanding of the world,
0:53:41 > 0:53:43and what things actually mean.
0:53:44 > 0:53:46Just to be very clear,
0:53:46 > 0:53:49this is this little robot learning to experience the world around him,
0:53:49 > 0:53:53to understand more, to have a greater potential of...
0:53:54 > 0:53:56- Yes.- ..cognitive capabilities.
0:53:59 > 0:54:01Where do you see robotics, like iCub,
0:54:01 > 0:54:03in a generation or two generations' time?
0:54:03 > 0:54:06Will iCub have grown up and gone to university,
0:54:06 > 0:54:08and gone and learned about the world around him?
0:54:08 > 0:54:10I can see this in the longer term,
0:54:10 > 0:54:13I don't see this happening in the next five to ten years.
0:54:13 > 0:54:16We talk about this happening in 20, 30 years' time,
0:54:16 > 0:54:17and it seems a long while off.
0:54:17 > 0:54:21We've been evolving for tens of millions of years,
0:54:21 > 0:54:22and you've got this little entity
0:54:22 > 0:54:25that's learning about the world around it now,
0:54:25 > 0:54:27and it's going from a blank slate
0:54:27 > 0:54:31to this seeing, interactive, responsive little unit.
0:54:31 > 0:54:33I think that's both the exciting point
0:54:33 > 0:54:35and the scary point with robotics.
0:54:35 > 0:54:37At the same time, we are in control,
0:54:37 > 0:54:40so we are determining the evolution of these systems.
0:54:40 > 0:54:42For now.
0:54:50 > 0:54:54On this journey, we've met some incredible robots.
0:54:54 > 0:54:59They're preparing for a voyage to Mars,
0:54:59 > 0:55:01becoming our friends and companions...
0:55:01 > 0:55:04My feet are off, my hands are off.
0:55:04 > 0:55:07..navigating us through a chaotic world...
0:55:08 > 0:55:12..and some are even able to learn like us.
0:55:12 > 0:55:17- Learn "ball".- For me, this is the most exciting time.
0:55:17 > 0:55:18I like to learn.
0:55:18 > 0:55:20This is a ball.
0:55:20 > 0:55:22We are living right at the moment
0:55:22 > 0:55:26when robots start to gradually piece things together,
0:55:26 > 0:55:28the first tiny scraps of meaning,
0:55:28 > 0:55:32to create their own unique understanding of the world
0:55:32 > 0:55:33and themselves.
0:55:33 > 0:55:35He wants the cup.
0:55:35 > 0:55:40Once they've achieved this, we will be on the brink of a new era.
0:55:41 > 0:55:45There is no doubt that robots will continue to evolve
0:55:45 > 0:55:49and become more and more intelligent and that, one day,
0:55:49 > 0:55:52it just might be possible for them to develop consciousness.
0:55:54 > 0:55:58Imagine a robot that could feel the way I feel,
0:55:58 > 0:56:01that could be moved by strong emotion,
0:56:01 > 0:56:05that could love the way that I love my daughter.
0:56:06 > 0:56:07Wouldn't that be incredible?
0:56:11 > 0:56:13When I started this journey,
0:56:13 > 0:56:18my main concern was that if robots could develop minds of their own,
0:56:18 > 0:56:20they might become a threat.
0:56:23 > 0:56:27But now I've started to spend more time with robots,
0:56:27 > 0:56:29I do feel like I can trust them.
0:56:29 > 0:56:31It's responding easily.
0:56:33 > 0:56:36And if robots really could one day become conscious,
0:56:36 > 0:56:40we need to think not just about how they might affect us,
0:56:40 > 0:56:43but how WE could affect THEM.
0:56:45 > 0:56:47But perhaps my biggest fear right now,
0:56:47 > 0:56:50as we progress towards conscious machines,
0:56:50 > 0:56:53is not what we will need to do for robots,
0:56:53 > 0:56:56but what we will discover about ourselves.
0:56:58 > 0:57:00The whole of our society, our law,
0:57:00 > 0:57:02our education is based around consciousness,
0:57:02 > 0:57:04making conscious decisions.
0:57:04 > 0:57:07And if we show that, well, actually, that's quite trivial
0:57:07 > 0:57:10and we can reproduce it in an afternoon in a lab,
0:57:10 > 0:57:12then it's going to make people think,
0:57:12 > 0:57:15"Well, how important is human life because it is conscious?"
0:57:15 > 0:57:18Ultimately, the rewards will be positive,
0:57:18 > 0:57:20but you have to be very, very careful.
0:57:20 > 0:57:23Socially, it might be disruptive.
0:57:27 > 0:57:30The extraordinarily fast evolution of robots
0:57:30 > 0:57:34really is going to change our place in the world,
0:57:34 > 0:57:38and that raises urgent social issues for us all.
0:57:39 > 0:57:43We need to be responsible, to make sure that we stay in control.
0:57:45 > 0:57:47We have the opportunity right now
0:57:47 > 0:57:50to prepare for conscious robots
0:57:50 > 0:57:53that think and feel in the same way we do,
0:57:53 > 0:57:57to prepare for what I think is the inevitable.
0:58:02 > 0:58:05Investigate the past, present and future of robots
0:58:05 > 0:58:07and their effects on our lives.
0:58:07 > 0:58:09Go to the address on screen
0:58:09 > 0:58:11and follow the links to the Open University.