04/02/2017 Click


Download Subtitles

SRT

ASS


04/02/2017

Can a human give artificial intelligence a real live voice? Plus a trip to Tokyo with games designer Hideo Kojima and the latest tech gadgets at the London Toy Fair.


Similar Content

Browse content similar to 04/02/2017. Check below for episodes and series from the same categories and more!

Transcript


LineFromTo

Much to look forward to today on BBC One for the rugby, that is all the

:00:00.:00:00.

sport for now, now it is time for Click.

:00:00.:00:00.

sport for now, now it is time for Click.

:00:00.:00:11.

This week, meet our youngest reporter ever.

:00:12.:00:12.

We give a games legend something to play with.

:00:13.:00:18.

For decades, scientists all around the world have been trying to

:00:19.:00:48.

create a machine that can match our intelligence.

:00:49.:00:52.

And nowadays artificially intelligent algorithms

:00:53.:00:56.

can perform many tasks much better than us.

:00:57.:00:58.

For a long time scientists have been the use in games like

:00:59.:01:01.

chess, drafts and go as a benchmark for testing AI.

:01:02.:01:05.

And that's because all these games have a certain

:01:06.:01:07.

amount of unpredictability built into them.

:01:08.:01:12.

But this week the AI community has been celebrating a big

:01:13.:01:19.

win after a poker playing algorithm called Liberatus defeated four top

:01:20.:01:26.

human players in a 20 day match of heads up no limit Texas hold 'em

:01:27.:01:30.

I've been using poker as a benchmark for 12 years.

:01:31.:01:33.

Now the best AI has surpassed the level of the

:01:34.:01:36.

best humans in the strategic reasoning under imperfect

:01:37.:01:37.

But even at this big win is only a little step towards

:01:38.:01:43.

Intelligence one capable of sophisticated thought across a wide

:01:44.:01:46.

spectrum of areas, and solving problems just as well as a human

:01:47.:01:49.

It's a hard thing to think through, and

:01:50.:02:00.

But it's, I think it's impossible to forecast accurately.

:02:01.:02:11.

Speech has been another big challenge for AI

:02:12.:02:18.

Personal assistants and chat bots are becoming more

:02:19.:02:21.

sophisticated, but they so far can't fool us into thinking that they're

:02:22.:02:24.

But what if you thought you were talking to another person?

:02:25.:02:30.

Would that make you more likely to trust it?

:02:31.:02:39.

Well, two researchers at the London School of

:02:40.:02:42.

economics came up with an experiment to see if we would communicate

:02:43.:02:45.

better with AI if its messages were delivered to us by a human.

:02:46.:02:48.

They call this computer human hybrid the

:02:49.:02:50.

And to explore the concept, Jane Copestick found

:02:51.:03:06.

herself becoming an Echoborg herself.

:03:07.:03:07.

The Echoborg was inspired by research from Stanley Milgram.

:03:08.:03:09.

He is the Professor behind the controversial experiments on

:03:10.:03:18.

obedience in the 1960s, to see if people would deliver

:03:19.:03:21.

electric shocks to others if instructed to buy an

:03:22.:03:23.

Milgram also studied body perception, to

:03:24.:03:25.

determine if we hold preformed opinions

:03:26.:03:27.

of other people based on

:03:28.:03:28.

By using hidden earpieces, people could speak

:03:29.:03:31.

someone else's thoughts through their own body.

:03:32.:03:32.

The Echoborg has updated this research for the 21st

:03:33.:03:34.

century, to see if people will react better to artificial intelligence.

:03:35.:03:39.

Such as the messages from an online chat bot.

:03:40.:03:41.

If they are being delivered by a human.

:03:42.:03:44.

I'm in the first stages of testing this out by

:03:45.:03:47.

I'm starting my speech shadowing practice.

:03:48.:03:52.

The first step in becoming a fully fledged Echoborg.

:03:53.:03:56.

The professors have told me this process

:03:57.:03:58.

will take at least eight hours for me to get any good at it.

:03:59.:04:01.

I'm starting my first practice with JK

:04:02.:04:03.

Rowling's Harvard commencement speech.

:04:04.:04:04.

Members of the Harvard Corporation and the board of

:04:05.:04:07.

By shadowing speech, I should be able to quickly repeat

:04:08.:04:10.

back the messages from a chat bot so people won't realise

:04:11.:04:13.

It may seem something paradox, but there's horses in the

:04:14.:04:18.

I did something and scuttled somewhere.

:04:19.:04:25.

Now, to put it to the test, I'm meeting creator Professor

:04:26.:04:28.

Alex Gillespie at the London School of Economics.

:04:29.:04:32.

And Kevin Corti, who called in on Skype.

:04:33.:04:36.

Kevin is using a chat bot called Rose, which is not preprogrammed.

:04:37.:04:46.

The most noticeable problem in becoming a convincing AI are the

:04:47.:04:49.

delays while Rose thinks of a response to the question.

:04:50.:04:55.

I thought for a moment you might be a

:04:56.:05:19.

Republic of Ireland and Croatia and France.

:05:20.:05:22.

A magical place full of rain and crazy people.

:05:23.:05:26.

What you notice, they tend to be quite

:05:27.:05:40.

It takes each sentence as a stand-alone sentence.

:05:41.:05:43.

Some of them will speak like they are

:05:44.:05:45.

artificial intelligence, and some of them will pretend not to be.

:05:46.:05:48.

But although last time I spoke to which

:05:49.:05:51.

it said it was artificial intelligence.

:05:52.:05:53.

Our final test for the Echoborg was to

:05:54.:06:00.

bring it on stage in front of an audience of 700 people at the BBC

:06:01.:06:04.

What a lot of humans find difficult...

:06:05.:06:07.

How do I know you are human, how do you know I'm human?

:06:08.:06:41.

In fact, some of the audience members

:06:42.:06:52.

One thought it was a real conversation with a human, not

:06:53.:06:56.

Some people thought you didn't want to talk about

:06:57.:07:02.

That you were trying to avoid the question,

:07:03.:07:07.

they really thought you were trying to avoid the questions.

:07:08.:07:10.

Someone even said, had it been a man would it

:07:11.:07:12.

Without becoming fully fledged Echoborgs, we are already

:07:13.:07:19.

giving a voice to artificial intelligence everyday.

:07:20.:07:21.

Through the algorithms guiding our news

:07:22.:07:22.

consumption, to our shopping habits and online searches.

:07:23.:07:24.

We're bringing AI to life more and more.

:07:25.:07:29.

Projects like the Echoborg let us reflect on

:07:30.:07:31.

what this means for our AI future and perhaps even what it means to be

:07:32.:07:34.

Hello and welcome to The Week in Tech.

:07:35.:07:47.

It was the week that Facebook lost $500 million in a lawsuit.

:07:48.:07:52.

The case centres around the creation of

:07:53.:08:02.

the oculus rift virtual reality headset.

:08:03.:08:03.

The US Court ordered the payment after a jury found Facebook

:08:04.:08:06.

owned VR outfit Oculus used computer code belonging to ZeniMax, a media

:08:07.:08:08.

company which has a subsidiary which produces the video game Doom.

:08:09.:08:11.

They say you shouldn't cry over spilt milk.

:08:12.:08:17.

Online supermarket Ocado is testing a robot hand that can pack

:08:18.:08:22.

fruits and vegetables without damaging them.

:08:23.:08:24.

At the moment, human beings pack more fragile items, like

:08:25.:08:31.

But it's not just fragile foodstuffs feeling the pinch

:08:32.:08:38.

Researchers at MIT have created a claw made from

:08:39.:08:42.

hydrogel, that can pick up a live fish without causing it any harm.

:08:43.:08:45.

Sunday the team hopes the eellike robot can be used to help with

:08:46.:08:49.

Next, forget the selfie stick, so 2015, it's all about the

:08:50.:08:58.

Currently being crowd funded, the air selfie

:08:59.:09:05.

is a portable flying camera built into a mobile phone cover.

:09:06.:09:08.

And, as it's carried around with your mobile

:09:09.:09:12.

phone, never miss an opportunity for Internet narcissism ever again.

:09:13.:09:25.

If you're a fan of Metal Gear Solid, you might also be

:09:26.:09:28.

Considered the father of the stealth game genre.

:09:29.:09:32.

The Metal Gear franchise was a success at

:09:33.:09:34.

least in part thanks to his leadership.

:09:35.:09:36.

But now he's working on a new game called Death Stranding,

:09:37.:09:42.

which he showed to the world at the E3 video games

:09:43.:09:48.

We sent Stefan Powell, ace radio one news beat reporter,

:09:49.:09:59.

to meet Hideo Kojima in Japan and get an exclusive tour

:10:00.:10:01.

We're on our way to the studio now and

:10:02.:10:09.

it's been just over a year since he left Konami

:10:10.:10:12.

And we don't really know what he's been doing in that time.

:10:13.:10:18.

We know a little bit about his new project,

:10:19.:10:20.

Death Stranding, that's coming to the PlayStation

:10:21.:10:24.

Hopefully we get to find out a little bit

:10:25.:10:35.

glimpse into the future and what's come as well.

:10:36.:10:40.

Before that, though, there's the traditional gift

:10:41.:10:42.

I mean, what are you supposed to get a man

:10:43.:10:54.

who stood in front of a cabinet full of lifetime achievement awards?

:10:55.:10:56.

And I hear you are a bit of a Lego fan.

:10:57.:11:05.

Kojima isn't your average game designer.

:11:06.:11:09.

And this isn't your average office, either.

:11:10.:11:10.

Or your average company mascot, for that

:11:11.:11:15.

The man credited with changing the way many people approached game

:11:16.:11:22.

design is not taking his new venture lightly.

:11:23.:11:28.

He wants his next step is to be just as successful as his

:11:29.:11:31.

Clearing his mind of some of the negativity of recent years.

:11:32.:11:36.

Focusing instead on the future, new titles, new projects, and new ideas.

:11:37.:11:41.

TRANSLATION: I worked at my previous company

:11:42.:11:43.

for 30 years, and gained a

:11:44.:11:45.

But technology improves, the games market and the

:11:46.:11:49.

But what I do best, making games, does not really

:11:50.:11:56.

change, so I'm not worried about embarking on this new journey.

:11:57.:12:00.

The studio itself is pretty small, but

:12:01.:12:02.

has everything Kojima and his team need to crack on with the first

:12:03.:12:05.

The PlayStation 4 exclusive, Death Stranding.

:12:06.:12:13.

Details about which are still top secret.

:12:14.:12:15.

But whatever it turned out to be, he's

:12:16.:12:18.

TRANSLATION: We want this game to be something

:12:19.:12:26.

people can get into easily, but after they play for an hour or two

:12:27.:12:32.

they start to notice something a little different.

:12:33.:12:37.

It's something they haven't played before.

:12:38.:12:39.

Whenever I create something new, some people

:12:40.:12:40.

For example, when I first created a stealth game some people really

:12:41.:12:45.

wanted to just fight, so they didn't really like it.

:12:46.:12:49.

I want to create an experience that has the same effect

:12:50.:12:51.

The building up of his own studio is also a source of

:12:52.:12:55.

A journey that has been far more difficult than many would

:12:56.:12:59.

This tiny room was Kojima production's first office.

:13:00.:13:07.

Here he spent time not only designing Death

:13:08.:13:09.

Stranding, but refining his next big idea to change gaming as we know it.

:13:10.:13:15.

TRANSLATION: The way I see the future of

:13:16.:13:18.

gaming, think of it as

:13:19.:13:20.

All meshed together to create one type of

:13:21.:13:23.

The whistle-stop tour of Tokyo continues.

:13:24.:13:28.

Seeing the places that came between that first

:13:29.:13:32.

tiny room and the shiny new studio housing Kojima Productions today.

:13:33.:13:37.

It's a vision that has grown from a long love of technology.

:13:38.:13:40.

His association with Sony so important

:13:41.:13:45.

to the future of his new company is not something new, though.

:13:46.:13:48.

Here at their big tech exhibition in downtown Tokyo, he explains how

:13:49.:13:51.

technology of the past has had such a big impact on him.

:13:52.:13:54.

He says the first Metal Gear was made on this device.

:13:55.:14:03.

Looking back over the past raises questions of

:14:04.:14:05.

VR is often said to be the next big thing in gaming.

:14:06.:14:17.

Now this isn't VR like we know it now.

:14:18.:14:19.

But for Kojima, it's not so clear-cut.

:14:20.:14:22.

Do you think the games out there for VR at the moment are good

:14:23.:14:25.

enough to really sort of get the audience excited?

:14:26.:14:29.

TRANSLATION: It's easy in VR to try and do something

:14:30.:14:32.

scary, something from a high place, something erotic.

:14:33.:14:43.

But I think there's something beyond that.

:14:44.:14:45.

You can give people emotions that they haven't

:14:46.:14:47.

And can you tell us any of your ideas?

:14:48.:14:50.

So it looks like a Kojima virtual reality

:14:51.:14:52.

experience is not so far away, even if he will chair

:14:53.:14:56.

experience is not so far away, even if he won't share

:14:57.:14:59.

More proof that his appetite for making things is not on the

:15:00.:15:03.

The most revealing thing I've found during my

:15:04.:15:07.

time with Hideo Kojima is that still really

:15:08.:15:08.

time with Hideo Kojima is that he's still really

:15:09.:15:11.

passionate and enthusiastic about tech and gaming.

:15:12.:15:12.

He's got no plans to retire any time soon.

:15:13.:15:15.

In fact, he set himself a big challenge.

:15:16.:15:16.

He's changed gaming once, and now he plans doing it

:15:17.:15:20.

Playtime was never like this in my day.

:15:21.:15:34.

I've been taking a look at some of the latest toys hoping to light

:15:35.:15:38.

up the faces of children and grown-ups.

:15:39.:15:44.

And, inevitably, a few of them could be found at London's toy

:15:45.:15:46.

This looks like a drone in a cage and that's because it is.

:15:47.:15:55.

It's also a proof of concept for a toy

:15:56.:15:58.

that's going to be available later this year.

:15:59.:16:01.

Its inventor here is wearing this glove, which means you

:16:02.:16:03.

It all looks pretty simple, but I know

:16:04.:16:23.

you've been studying robotics for 15 years,

:16:24.:16:24.

so there's quite a bit more to this than meets the eye, isn't

:16:25.:16:27.

Once the science of gestures has been

:16:28.:16:31.

codified, and that's what we've been able to do,

:16:32.:16:33.

as you can imagine, we

:16:34.:16:34.

can bring all sorts of robotic toys out, and consumer devices.

:16:35.:16:37.

The brain itself is in the glove, in the

:16:38.:16:41.

And the algorithms embedded in the glove.

:16:42.:16:43.

The drone is merely a conduit for the gestures being

:16:44.:16:46.

There was also a clear trend towards giving kids a deeper level

:16:47.:16:50.

of control when it comes to toy gadgets.

:16:51.:16:56.

This is a robot that aims to help kids learn to code.

:16:57.:17:02.

They can operate it manually through the App,

:17:03.:17:08.

or set up sequences of the functions they'd like it to carry out.

:17:09.:17:11.

It looks pretty raw when you've got all

:17:12.:17:13.

these leads and buttons, so it really is giving kids a chance to

:17:14.:17:23.

I also recently got my hands on a drone that kids can

:17:24.:17:26.

programme, spending time tweaking code at a computer or using drag and

:17:27.:17:29.

I had a play around with some of the drone's functions.

:17:30.:17:33.

So maybe that shows who the real kid is.

:17:34.:17:35.

First of all I press W, which should get the drone up and running.

:17:36.:17:38.

This is a spot of that well-known activity, drone bowling.

:17:39.:17:41.

Yes, the skittles are down here on the floor.

:17:42.:17:56.

It's not just about flying, though, you

:17:57.:18:00.

To do that, you swap the wings for wheels.

:18:01.:18:04.

Last year we learned quite how much of an appetite there was

:18:05.:18:07.

And give the big kids a chance for some

:18:08.:18:26.

This gaming robot, much like virtual avatars,

:18:27.:18:47.

It's also customisable and upgradable, with the ability to add

:18:48.:18:53.

wheels or even take on another robot in the room.

:18:54.:18:57.

Or if you want to get yourself moving, how about a personal

:18:58.:18:59.

This prototype has limited functionality, but still

:19:00.:19:20.

Not that it fought too hard when I decided I'd had enough.

:19:21.:19:32.

Now, if you're a Cinefile, you know we

:19:33.:19:34.

have officially entered awards season.

:19:35.:19:38.

Yes, the red carpets, the speeches and the campaigning have

:19:39.:19:42.

Well, this year at Click, we've decided to give those

:19:43.:19:50.

wonderful magicians behind the camera, namely the visual

:19:51.:19:52.

effects artists, their proper due with a

:19:53.:19:53.

series of exclusive features on some of the most

:19:54.:19:56.

memorable films from the

:19:57.:19:57.

First up, a return to the wizarding world with BAFTA nominee

:19:58.:20:01.

Fantastic Beasts and Where to Find Them.

:20:02.:20:05.

Mr Scamander, do you know anything about the wizarding

:20:06.:20:07.

The earliest Potter film I worked on was the second

:20:08.:20:15.

Then went on to work on subsequent films for production.

:20:16.:20:20.

The big difference, I would say, now,

:20:21.:20:22.

doing Fantastic Beasts, for instance, was we were doing creature

:20:23.:20:26.

design in the computer from day one.

:20:27.:20:28.

We were animating creatures, showing David what they looked like.

:20:29.:20:32.

And getting into a developmental study through

:20:33.:20:40.

Just something we could never have done before so

:20:41.:20:44.

When we decided we didn't like it we could modify it and

:20:45.:20:48.

change it very quickly on the computer.

:20:49.:20:50.

We would model something up, in ZBrush, then very quickly

:20:51.:20:55.

think, that looks cool, let's stick a rig in it.

:20:56.:20:57.

So, sticking essentially the bones in to be able

:20:58.:21:00.

And we had quite worked up animation studies of

:21:01.:21:07.

a lot of stuff that just didn't make the film.

:21:08.:21:09.

We made simple models of the creatures, then brought in some

:21:10.:21:12.

Some of them were actually, the team who did

:21:13.:21:18.

Warhorse for instance, then we had a full-size erumpent puppet,

:21:19.:21:20.

And they were able to use that onset.

:21:21.:21:27.

And our guide was always what we've done in

:21:28.:21:30.

the computer in terms of the animated previews.

:21:31.:21:33.

It meant we could put something on set for Eddie

:21:34.:21:35.

Redmayne to react to or perform against.

:21:36.:21:38.

So the erumpent, 17 foot tall, 20 foot long, was able to be

:21:39.:21:42.

onset, and everybody could see how big it

:21:43.:21:44.

was and where she was on the

:21:45.:21:46.

set, then we could frame the camera for her and Eddie could play against

:21:47.:21:49.

One of the key things about this film was that you believe the

:21:50.:21:54.

If you believe the actor, and you believe the

:21:55.:21:58.

creature's there, that's what makes it work.

:21:59.:21:59.

When you see an actor and the eyeline isn't right, and

:22:00.:22:02.

the creature doesn't seem to be responding to them, you know there

:22:03.:22:05.

The niffler, ultimately, was a fairly tough

:22:06.:22:09.

A lot more close-up, I guess, than you would have done, for

:22:10.:22:16.

a lot longer than you would have done a few years ago.

:22:17.:22:19.

You look at the niffler and what makes him so

:22:20.:22:21.

animalistic and real is all that small breathing, all the secondary

:22:22.:22:24.

stuff, it is in the overall performance, it is all that

:22:25.:22:26.

That we put loads of work into, that you

:22:27.:22:30.

kind of think, God, really, do you notice?

:22:31.:22:31.

And it's like, no, you don't notice directly, but you do notice

:22:32.:22:34.

because we all look at human beings all the time.

:22:35.:22:37.

You've got a pretty big price on your head, Mr

:22:38.:22:51.

Gnarlak, our goblin, was, you know, a really

:22:52.:23:00.

Probably the most advanced digital humanoid type character I think

:23:01.:23:06.

I think Ron was fantastic to work with, he wore a

:23:07.:23:12.

performance patch, a headset, so he had

:23:13.:23:14.

17 facial markers on, doing the

:23:15.:23:16.

facial capture meant in our post-stage we were able to deliver

:23:17.:23:21.

That was just the pure capture applied to

:23:22.:23:24.

It was a massive leap in shading and the technology driving

:23:25.:23:32.

Whilst there is the animation, there are also all the

:23:33.:23:35.

systems we've created to help drive muscles, skin,

:23:36.:23:37.

All of that stuff's working up and up.

:23:38.:23:46.

Every film, we're pushing evermore, trying to get to that

:23:47.:23:50.

More Oscar hopefuls and special perfect

:23:51.:23:56.

Follow us on Twitter throughout the week.

:23:57.:24:02.

Thanks for watching and we'll see you soon.

:24:03.:24:17.