Your Face or Mine Click

Now on BBC News,

it's time for Click.

This week, lady becomes model.

Man becomes ape.

And Spen becomes Trump.

AS DONALD TRUMP:

Great, the best.

OK, movie quiz time.

Five points if you can

name this film.

Correct - it's Raiders

of the Lost Ark.

No, that is not Harrison Ford,

that is the face of Nicholas Cage.

OK, try this one.

Yes, it is The Fellowship

of the Ring.

100 points if you spotted

Nicholas Cage, Nicolas Cage

and Nicholas Cage.

So, what on Earth is going on?

We're just about getting used

to the idea that there

are loads of fakes online.

Fake news, fake tweets,

fake photoshopped images.

But these videos are a whole

level above anything

that we've seen before,

and they may have consequences that

go far beyond just switching

out a few movie stars.

A lot of what we talk about over

the dinner table is,

we live in a diverse world.

Researchers at the University

of Washington released

this video last year,

which used a computer vision

algorithm to very convincingly

doctor Obama's mouth movements

to make him lipsync to something

he said in a different interview.

A lot of kids, the doors that

have been opened to me

aren't open to them.

And with the tricks and tools

of machine learning becoming

better and easier to use,

it's now possible to do this without

a particularly powerful computer.

Remember the Nick Cage

videos from earlier?

Well, this mix of Donald Trump

and Angela Merkel was created

using the same tool,

a tool called Deepfakes.

To be clear, this is not just a face

swap like you might see on Snapchat.

This is artificial intelligence that

has learned what Trump's face looks

like and then made it copy

Merkel's facial expressions.

What's fascinating is that these

weren't made by a team

of researchers, or a Hollywood

visual effects department.

These were made by individuals

following an online tutorial

on a desktop machine.

Now, to see how easy it is,

we're going to do it.

We're going to take my face

and make me President.

We trained a neural network

by feeding it video of some

of my past appearances.

We mixed it with President Trump's

State of the Union Address.

The software broke the video

into individual frames,

run them through the network and,

in less than a day,

this was the result.

All of us, together, as one team...

So, this is the original

video of Trump.

And this is me, on his head.

We all share the same home.

I'm not sure it's an improvement,

but that does seem to be

President Spenley Trump.

The other half of the experiment

didn't go quite so well.

This is Click presenter

Donald Kelly.

Now, this was a very short

and quick experiment.

It's far from perfect.

It's blurry, you can see

the edges and sometimes -

well, it's just downright scary.

But had we left the network to train

for longer, on better videos,

we could have got much

more convincing results.

Now, Deepfakes has hit

the headlines in recent weeks,

but it's not because of Trump

or Nicolas Cage, or even me.

It's because of porn.

An online community has been

using AI to put Hollywood

actresses' faces onto adult

movie stars' bodies.

Now, obviously we can't show

you much of the resulting videos,

but they feature a number

of female celebrities.

The videos are stolen,

and so are the faces.

I was surprised but,

at the same time, not really,

because of how technology

is advancing so much that it's just

seems kind of inevitable.

However, the context

of the material, taking a celebrity

or a well-known face -

or even anyone's face,

for that matter - and putting it

on a porn performer's body,

and essentially creating this

fake, non-consensual

porn, it's disturbing.

It was definitely

a disturbing thing to see.

Are people going to take, like,

pictures of ex-girlfriends and put

them on porn performers?

Issues surrounding consent

and copyright have led the website

hosting this content to ban Deepfake

photography, and the

communities making it.

Reddit too, the site

where the community started,

changed its rules to forbid

involuntary fake pornography.

Away from porn, doesn't take much

imagination to see how one

could create international outrage

by making fake statements

from world leaders.

Something that may become possible

very soon, thanks to some software

that we looked at last year.

This is Lyrebird.

The idea here is that I can train

a neural network with samples

of my voice and then it will be able

to speak like me.

Harry hoped he would see some

success from the current project.

Parents should look out for...

The software asks you to read out

at least 30 sentences

of its choosing, from which it can

pull out the basic building blocks

of words, the phonemes,

that can then be put back together

in any order.

In other words, "in other words".

I've always been a big

fan of One Direction.

They were, quite frankly,

better than the Beatles.

SPENCER LAUGHS.

Although the creators of Lyrebird

are aware that this technology

could be misused, they say that

by releasing it as a free tool,

well, at least the public

will become aware that fake voices

are already a reality.

AS DONALD TRUMP:

Great, the best!

One idea that we are considering is

to watermark the audio

samples that we produce.

So we are able to detect immediately

if it is generated by us.

So, how do we protect ourselves

from having our online photos,

videos and sound recordings used

to create fake us-es?

At the moment, we are in a wild,

wild west situation.

We don't know the attitude

of the courts to this problem.

We don't have a clear piece

of legislation that would cover it.

We have piecemeal laws on privacy,

copyright, trademark and passing off

that would be useful to somebody

in trying to stop

this from happening.

We don't have a clear legal

definition and we don't have a clear

piece of legislation

that is exactly on point.

And until we have that, this legal

uncertainty will continue.

The morality and the legality

of Deepfakes are murky issues.

Just as we are wrestling

with the fact that we can't trust

what we read, very soon we will need

to confront the fact

that we can't trust anything

we see or hear either.

Hello, and welcome

to The Week In Tech.

It was the week that Google's

dedicated health farm announced it

has created an algorithm that can

protect high blood pressure,

as well as risk of heart attacks

or strokes, simply by scanning

a person's eyes.

As Boston Dynamics keep

on showing us how intelligent

robots are becoming,

experts this week warned that

artificial intelligence in the wrong

hands is ripe for exploitation.

Drones turned into missiles,

fake videos manipulating public

opinion and automated hacking

are just three of the

And Amazon CEO Jeff Bezos

is getting a new clock.

But it's not one you can

pick up from a shop.

Costing a huge $42 million -

or £30 million - it is designed

to run for 10,000 years

and construction began this week

inside a hollowed-out

mountain in Texas.

The Vice President of Facebook's

adverts found himself

in hot water this week,

after tweeting that Russian-backed

ads were not designed

to sway the US elections.

Rob Goldman's tweet came after 13

Russians were charged with meddling

in the election via social media.

His views were not those

of Facebook itself.

And a conspiracy theory video

claiming that survivors of last

week's Florida shooting are "crisis

actors" somehow became YouTube's

number one trending clip

before being removed.

And finally, pictures

of Samsung's latest phone,

the Galaxy S9 were leaked ahead

of its launch at the

Mobile World Congress,

starting on Sunday -

ironically, via an app released

by the company itself.

We'll move on now

to video game news.

Remember Nintendo's Switch,

its hugely successful console that's

both mobile and which plugs

into a TV?

Well, the Japanese gaming giant has

now created a host of rather unusual

new peripherals which wildly alter

how the machine is used.

And Marc Cieslak has been getting

all bent out of shape over it.

HE PLAYS A MUSICAL SCALE.

You may be forgiven for thinking

that this cardboard is the packaging

for the new peripherals

for the Nintendo Switch console.

However, the cardboard

are the peripherals themselves.

Called Labo, it's a range of devices

which includes things like a piano,

motor bike handlebars,

fishing rod and even a robot suit.

Straps on the shoes...

I might look like I'm stomping

around in a slightly weird way.

But this game asks you to really get

into the character of a giant robot.

And, if I pull out my visor...

I activate first-person mode.

For precision destruction!

Called Toy-Cons, they are all

constructed from folded cardboard.

Some use elastic bands

and all use the Switch's motion

sensing controllers.

I think Labo is a big deal

for the Nintendo Switch,

just because it proves that Nintendo

is capable of continue to innovate

on an already innovative product.

The fact that it is made out

of cardboard and your existing

controllers fit in, I think that

will blow parents' minds and,

more importantly, blow

children's minds as well.

But before you can play

with your Toy-Con, you've

got to build at first,

something that you might worry

requires the prowess of an origami

expert crossed with the advanced

flatpack furniture building skills

of a self-assembly sensei.

Building these devices takes

varying lengths of time.

More complicated Toy-Cons,

like the robot suit,

can take up to eight

hours to complete.

But that's part of the appeal

of Labo, taking pleasure

from the building of the devices

that you are about to

use, and understanding

how they go together.

A little bit of patience and some

deft folding results in this.

Nintendo reckons this

is a radio controlled car.

Last time I looked, cars had wheels.

My completed Toy-Con,

which I can make move around,

because the Switch controllers have

got HD rumble and it

means that you can have

different levels of rumble.

Allowing this particular

Toy-Con to move about.

Now, this only took me about ten

minutes to build, all in all.

And because the Switch controllers

have an IR camera in them,

on the Switch itself,

you can see where

the Toy-Con is going.

Each one of the Toy-Cons

comes with a game.

Some are more complicated

than others, but will require

an element of physical control,

which comes courtesy

of the folded cardboard.

The games themselves

are more like mini-games.

But that's not the point.

This is more about creativity

and making something than it is

a hard-core gaming experience.

But I do question the durability

of cardboard peripherals.

How does that go back in there?

Not very, based

on my time with them.

We've managed to have a pit

stop with our very own

cardboard mechanic.

So, while I managed to damage my

cardboard motorcycle,

repairs are really quite easy.

There were two different offerings

so far, the Variety Pack,

which includes five different

Toy-Cons, priced at £59.99, and

the Robosuit, which costs £69.99.

That seems like a lot

of money for cardboard toys

with bits of string for guts.

Nintendo hasn't yet said

whether they are going to give

you replacement parts for that,

or whether you are going to have to

scavenge cardboard from supermarkets

or things like that.

So it's going to be interesting

to see how much Nintendo

are expecting you to spend on top

of the base game and cardboard kits.

This week, Caterpillar announced

the release of a new smartphone.

You'd be forgiven for not even

knowing they produced such a thing.

These devices are specifically aimed

at the construction industry.

But this one has a few

interesting features.

An upgrade to their FLIR

thermal imaging camera,

the addition of a laser beam

for measuring how far away

something is, or room size,

and the standout feature, a nose.

Yes, it can smell.

Or, more specifically,

it has an indoor air quality sensor

which aims to alert users

if there are high levels

of volatile organic compounds -

or VOCs - in the air,

something commonly found in paint,

solvents and cleaning products.

Sound a bit niche?

Well, its creators don't think so.

Builders, plumbers, electricians,

carpenters, farmers.

These type of people kind

of generally get overlooked

by the everyday phone vendors.

So, what we are doing

is understanding the technology

that we can integrate

into our products that really

makes their lives better.

Next week on the show, we'll be

bringing you all of the latest news

and releases from NWC in Barcelona.

That was Lara.

Now, earlier we looked

at how one face can be

transplanted onto another.

Next, it's time to meet the people

who took this man's face

and body and turned him

into a completely different species.

The team behind War For The Planet

Of The Apes is hoping to win

the Best Visual Effects Oscar next

weekend.

If it does, it will be the second

Oscar in a row for Dan Lemon.

Bad place.

Human zoo.

Bad human.

Bad humans.

Soldiers.

Soldiers...

War For The Planet Of The Apes

represents the effort of over 500

people in visual effects.

Some of us worked on the movie

for, like, two years.

The movie runs about two

hours and 20 minutes,

and all but 15 shots in the movie

had visual effects.

Many of the heaviest shots

in the movie took over 1400 hours

per frame to render.

So, that's a core hours.

So we have a lot of those processors

working at the same time.

Maybe the frames wouldn't come back

for two days, three days.

Bad place.

I find long time ago, after zoo...

So, the razor's edge

that we walk is figuring out how

to take our apes and adjust them,

and make the performance readable

to a human audience,

so it is legible as something that

you and I can connect with.

But also, not to take it so far

that suddenly it doesn't

feel ape-like any more.

Somehow it feels too human.

That is the fine line

that we ride and that

requires our animators to really...

You know, it's their kind

of skill and experience that

comes into play there.

This particular movie plays as sort

of this epic Western, you know?

Where they are travelling from one

location to the next and you get

to see a lot of different

environments, and spaces.

The big environment

was the prison camp.

That is where most of the second

half of the movie takes place.

The prison camp, we actually built

it near the Vancouver Airport

in British Columbia.

It was a 70,000 square feet set.

So it was a really big set.

But it was surrounded by over

100,000 square feet of green screen.

We needed all that screen

because we had big wide shots

that were set quite low.

And beyond this set, we needed

to put the Sierra Nevada mountains.

We used a process called

photogrammetry, where we went up

in a helicopter and shot reference

up and down the Sierra Nevadas

in that kind of region.

And then we turned those

still photographs into

three-dimensional geometry

using the photogrammetry process.

And we were able to use those pieces

of geometry sort of like a kit set.

And in the kind of traditional model

making sense, we were able to take

geometry from those mountains,

cut them up and then put

them together into a way

where we could achieve specific

compositions that matched what

the director was after for that set.

We developed this new piece

of software called Totara.

Totara is an ecosystem simulator.

So, we built the terrain

and we sort of map resources.

We said, this is where the good soil

is, this is where the bad soil is.

And then we let the plants

grow from these seeds.

We ran this similation over

the equivalent of 80 years

inside the computer.

And it allowed the trees to grow up.

And as the trees grew,

they competed with one

another for resources.

So, they would kind of try

to reach higher and higher.

And what happened was a lot

of complexity emerged

from that simulation.

It looked a lot more realistic,

and we got a lot more realistic

variation and naturalism

from that simulation.

One of the things that are so great

about our job is not knowing

what the next thing is.

And that, for us, is the thing

that is so much fun,

to be able to be, like,

OK, here is the creative challenge.

How can we take our technical tools

and bend them to tell this story,

or what can we invent,

what can we make up to be

able to tell the story?

No, no, no...

OK...

You might remember that late last

year I had a brilliant chat

with Andy Serkis about his role

as Caesar in War For The Planet

Of The Apes, and the art

of motion capture in general.

And you can watch that again right

now on our YouTube channel.

And next week we will get exclusive,

behind-the-scenes access to another

visual effects Oscar nominee,

Star Wars: The Last Jedi.

Now, as it happens, the visual

effects company that power Star Wars

has also been involved in something

rather different recently.

Just in time for Fashion Week

season, the London College

of Fashion has teamed up

with Industrial Light & Magic to put

on a show with a difference.

Two years in the making,

ILMxLAB is debuting its live CGX

technology in an augmented

reality catwalk experience.

Real-life models are joined

by a virtual avatar

that is controlled by a human,

wearing ILM's signature performance

capture gear backstage.

As the performer walks around,

a specially designed rig,

motion capture and depth sensing

cameras tracked her every move.

The avatar mirrors her

physicality in real-time.

Ahead of the show, we got

an exclusive sneaky

peek behind the scenes.

So, when we are calibrating,

we are solving your skeleton

from all the markers,

so that you know, OK,

that is how your wrist moves,

that is how your hip moves.

It gets the right

proportions and everything.

Having a live event,

where we are using the technology

that doesn't require a headset to be

put on, and it is also

not Star Wars-based,

that is quite a two-part

departure for us.

So, you know, we are really excited

to go into this new foray of what it

means to collaborate on a live

event, number one, and, number two,

just being out here in London.

It was really interesting to design

with the consideration

that it would be the Jedis.

So, the textures and materials

were chosen consciously

to make sure it reflects

what you guys can create.

No catwalk is complete, of course,

without a wardrobe change.

And so, during the show,

the virtual model's clothing

transforms via the real-life

performer taking a bow.

The gesture triggers

the change live.

To view the transformation, though,

tending fashionistas have to look

at a large five by three

metre LED screen.

It shows a live video feed

of the stage taken from the balcony.

The room around them comes to life,

too, with overlaid graphics,

turning this traditional interior

into a Macau inspired street scene.

The fashion industry

is based on emotion.

There is enormous

desire for newness.

So integrating technology

and fashion allows us

to create new experiences,

new garments, in this

case, for consumers

to enjoy and experience.

So, the last five minutes, all of

the models will stand in position.

Seeing how the fabric drapes

and falls, and how that mimics

into digital format,

that makes me really excited

because of the potential

of what you can do to fabric.

So, that kind of control

is something that I guess,

as designers, we don't

really normally have.

And that's what we really wanted

to kind of push for.

Eventually, the aim is to shrink

this elaborate setup down to a pair

of wearable goggles.

Let's hope they are

suitably en vogue!

Bravo!

And that's it from us for this week.

Don't forget we live on Twitter

@BBCClick and on Facebook, too.

You can find us there, hanging

around every day of every week.

Thanks for watching

and we will see you soon.