10/12/2016 Click


10/12/2016

Forget self-driving cars, Click takes to the skies in an autonomous aircraft. Plus, poverty-predicting software and sci-fi shopping.


Similar Content

Browse content similar to 10/12/2016. Check below for episodes and series from the same categories and more!

Transcript


LineFromTo

Now on BBC News it is time for Click.

:00:00.:00:00.

This week, mapping the poorest parts of the world.

:00:00.:00:07.

And, hands up who is not flying the plane?

:00:08.:00:37.

There are things happening in artificial intelligence right

:00:38.:00:46.

now that will fundamentally change our world.

:00:47.:00:51.

Soon, machines will learn to do our jobs.

:00:52.:00:56.

And at that point, things get very interesting.

:00:57.:01:08.

We'll talk more about the consequences of an automated

:01:09.:01:11.

society in a few minutes but, after thinking, walking and driving,

:01:12.:01:14.

have you ever wondered how hard it would be for a computer to fly?

:01:15.:01:24.

I'm not talking about drones that can fly between points, follow

:01:25.:01:28.

I'm talking about aircraft that could intelligently

:01:29.:01:35.

decide on a flight plan, just as a human would.

:01:36.:01:41.

And that is what Mark has been hunting down.

:01:42.:01:48.

Here at BAE Systems in Woolton in Lancashire, they are used

:01:49.:01:52.

They built and have tested Eurofighter Typhoons here.

:01:53.:02:00.

Today, however, I am going to take a flight in an aircraft that is much

:02:01.:02:04.

This is a Jetstream 31, a small passenger aircraft.

:02:05.:02:12.

It's a design from the 1980s but it's currently used by BAE

:02:13.:02:16.

Systems as a flying test-bed for technology which could lead

:02:17.:02:19.

Maureen McCue is head of research here.

:02:20.:02:27.

Very well flown and understood aircraft from the outside,

:02:28.:02:32.

but on the inside, it's filled with the latest technology.

:02:33.:02:37.

That technology will eventually allow this aircraft to fly

:02:38.:02:43.

Today, they are testing the plane's ability to detect and avoid

:02:44.:02:47.

clouds as well as testing its satellite communication systems.

:02:48.:02:52.

But take-off and landing will still be handled by human

:02:53.:02:55.

pilots and the plane will be remotely controlled

:02:56.:02:57.

How does this fit into the autonomous equation?

:02:58.:03:01.

At the moment, it's effectively a remotely controlled aircraft.

:03:02.:03:03.

It is, and really, with autonomous operations,

:03:04.:03:06.

you need to progressively expand the boundary.

:03:07.:03:09.

You can't start with such a big bang right out at the full range

:03:10.:03:13.

This humble looking outbuilding houses the grand station.

:03:14.:03:23.

Here, a pilot will remotely fly the plane and he can ensure it

:03:24.:03:27.

will react to instructions from air traffic control.

:03:28.:03:35.

I would expect to see a joystick and images coming

:03:36.:03:39.

through from the cockpit by you're not going to be flying

:03:40.:03:42.

No, everything is through the numbers that you can see there.

:03:43.:03:47.

These flights are taking place in uncongested airspace.

:03:48.:03:53.

Today, we will be flying over the Irish Sea.

:03:54.:03:59.

To help fly itself, this aircraft uses data from satellites

:04:00.:04:10.

as well as identifying radio signals broadcast by other aeroplanes, so it

:04:11.:04:13.

It is also fitted with a camera that can see other air

:04:14.:04:20.

users, even if there are not warning signals.

:04:21.:04:24.

So right now, the pilots aren't actually flying the aircraft.

:04:25.:04:26.

From that 15-year-old Dell laptop that's probably Windows XP.

:04:27.:04:34.

At this point, the aircraft is flying autonomously with a human

:04:35.:04:38.

Handing control of the computing over to the autopilot in the back

:04:39.:04:46.

and once established on the route, I can hand the computing

:04:47.:04:51.

over to Clive by the satellite on the ground.

:04:52.:04:57.

So that little shed-like building we were in earlier with Clive

:04:58.:05:00.

in front of the computer, he is now flying the aircraft.

:05:01.:05:04.

Over the course of the testing of this aircraft, it's going to have

:05:05.:05:09.

to perform a variety of different, complex tasks.

:05:10.:05:12.

For instance, it's going to have to recognise and avoid bad weather.

:05:13.:05:15.

Not just weather, but other aeroplanes, too.

:05:16.:05:19.

It will eventually be able to select a safe landing spot

:05:20.:05:22.

Today, we can't really test its weather detection

:05:23.:05:30.

abilities though as, unusually for the UK,

:05:31.:05:32.

BAE suggests that autonomous aircraft could be used

:05:33.:05:38.

to perform dirty, dangerous or repetitive tasks.

:05:39.:05:42.

But could this technology be introduced into

:05:43.:05:44.

At the moment all commercial aircraft have a set number of crew.

:05:45.:05:50.

There are programmes in existence looking at how you reduce crew

:05:51.:05:53.

either planned from the outset or, in the case of an emergency,

:05:54.:05:58.

you've got the autonomous system as a fallback so that you can

:05:59.:06:02.

still have perhaps a two-crewed aircraft on a certain length

:06:03.:06:05.

of flight but one of the crew happens to be an autonomous helper

:06:06.:06:09.

But what happens when things go wrong?

:06:10.:06:19.

While aerospace manufacturers are exploring the possibilities

:06:20.:06:31.

of fewer cockpit crew, what do commercial pilots think?

:06:32.:06:34.

To find out, we paid the British Airline Pilots Association a visit.

:06:35.:06:39.

Many decades of looking at aviation has brought us to the position

:06:40.:06:42.

where we have come to the conclusion it's best to have two

:06:43.:06:46.

pilots in the cockpit, because if you reduce that to one,

:06:47.:06:49.

the problem you've got then is you've got no one to cross check

:06:50.:06:52.

Take for example, the miracle on the Hudson.

:06:53.:06:55.

When the aircraft lost both its engines, the pilots had

:06:56.:07:06.

to have a discussion and they decided their only course

:07:07.:07:09.

No computer can be programmed to do that.

:07:10.:07:13.

The flight testing of autonomous aircraft continues but the debate

:07:14.:07:16.

about regulating them and how they are going to be used has

:07:17.:07:19.

That was Mark and this is Tim Harford, columnist

:07:20.:07:23.

for the Financial Times and Tim, you've written a book about how

:07:24.:07:26.

the systems that we now rely on can sometimes backfire.

:07:27.:07:32.

What do you make of the idea of planes that might only need one

:07:33.:07:36.

And of course, autopilots have made planes safer

:07:37.:07:46.

but what worries me is, what happens when

:07:47.:07:48.

No system is perfect, including a system where

:07:49.:07:51.

I guess when it goes wrong, it has two hand back

:07:52.:07:55.

Autopilot hands back to the human in the cockpit but then what?

:07:56.:08:00.

The human is out of practice, the human is not used to flying

:08:01.:08:03.

the plane and because the autopilot has failed, it's probably

:08:04.:08:06.

There is a worrying example of this a few years ago.

:08:07.:08:11.

An Air France crash over the Atlantic Ocean.

:08:12.:08:14.

The plane was flying quite high above a storm.

:08:15.:08:18.

The autopilot disconnected and the pilots just weren't used

:08:19.:08:23.

They were only used to operating the plane on take-off and landing

:08:24.:08:32.

and they flew a perfectly good plane into the Atlantic Ocean

:08:33.:08:35.

because they were confused about what was happening.

:08:36.:08:37.

They killed everybody on board, an absolute tragedy, and this

:08:38.:08:40.

The autopilots are normally so safe, so reliable,

:08:41.:08:45.

that when they fail, the pilots find

:08:46.:08:47.

I guess the next question is, what about autonomous cars?

:08:48.:08:53.

We have been talking about how they will blissfully drive us around

:08:54.:08:57.

I suppose for the foreseeable future, they won't be good enough

:08:58.:09:02.

I guess they'll never be 100% reliable but the model

:09:03.:09:10.

where if it's confused, it hands back to the human,

:09:11.:09:13.

You are there with your bagel, your coffee, your newspaper.

:09:14.:09:22.

You look up, there's a bus coming towards you, and the car goes,

:09:23.:09:25.

autopilot disengaged, human take control, it clearly not

:09:26.:09:28.

What makes more sense is for the human to be driving

:09:29.:09:32.

and for the computer to be watching out for a dangerous situation,

:09:33.:09:36.

for the computer to take over if there's a problem.

:09:37.:09:38.

Humans get bored, get distracted, lose their skills.

:09:39.:09:43.

None of these things happened to computers.

:09:44.:09:49.

I guess we are in an extended period of time before the far future

:09:50.:09:53.

happens and computers drive and fly us.

:09:54.:09:55.

We've got possibly decades if not a century of being in this

:09:56.:09:58.

interim period where, if there is a problem,

:09:59.:10:01.

we are going to end up blaming them for this really unusual,

:10:02.:10:04.

weird crash that a human wouldn't have made.

:10:05.:10:07.

Yes, and I think a glimpse of that is where we are asking

:10:08.:10:10.

the computers to make a decision not about planes or cars but about,

:10:11.:10:14.

for example, who'll get a promotion or who gets a special deal in a shop

:10:15.:10:18.

or who gets arrested for shoplifting because the computer

:10:19.:10:21.

We are already asking computers to make this sort of decision

:10:22.:10:26.

and the lesson of the paradox of automation is that we need to be

:10:27.:10:30.

much more savvy about the fact that computers do make mistakes

:10:31.:10:33.

Hello and welcome to The Week In Tech.

:10:34.:10:45.

It was the week that inventor Haiyan Zhang developed a smart

:10:46.:10:48.

wristband to help people with Parkinson's disease

:10:49.:10:50.

The device's in-built motors vibrate to distract

:10:51.:10:56.

It was also the week that we discovered queueing

:10:57.:11:02.

at the shops and using those beyond infuriating self-service

:11:03.:11:04.

checkouts could soon be a thing of the past.

:11:05.:11:08.

Amazon has unveiled a sci-fi store in Seattle that uses your smartphone

:11:09.:11:11.

and advanced technologies like deep learning, computer vision and sensor

:11:12.:11:16.

fusion to automatically detect when products are taken

:11:17.:11:20.

When you're done, you can simply trot off and then wait

:11:21.:11:26.

for that gargantuan virtual receipt to follow.

:11:27.:11:30.

And if you feel like you're forever stuck in traffic,

:11:31.:11:32.

Audi has rolled out an update to make every second count.

:11:33.:11:36.

Its new traffic light information feature tells drivers exactly how

:11:37.:11:39.

long they'll have to wait behind a red signal before it turns

:11:40.:11:42.

green and the length of time it will stay green.

:11:43.:11:47.

It works by connecting directly to the city's

:11:48.:11:50.

And finally, robotic research has reached new heights, literally.

:11:51.:11:56.

This hopping mad bot developed at UC Berkley cannot only jump a meter off

:11:57.:12:04.

the ground but can then again jump off objects to reach

:12:05.:12:07.

Inspired by the agility of bushbabies, researchers hope

:12:08.:12:12.

it can one day identify jumping spots for itself.

:12:13.:12:21.

Which these days isn't guaranteed to be true.

:12:22.:12:30.

In the run-up to the US election, for example, the Speaker

:12:31.:12:33.

of the House of Representatives did not get naked, the Pope did not

:12:34.:12:37.

endorse Donald Trump and he did not win the popular vote,

:12:38.:12:40.

but these stories, from websites posing as real news sites,

:12:41.:12:45.

Of course, it doesn't help that in 2016, the real news sounds

:12:46.:12:51.

But anyway, it's made events like the Trust Hack here in London

:12:52.:12:57.

Here, journalists and technologists from large news organisations

:12:58.:13:06.

are workshopping ways to help readers tell the difference

:13:07.:13:11.

between well-researched journalism, propaganda, advertising,

:13:12.:13:13.

The thought is to provide images like icons back-up materials

:13:14.:13:20.

that the public could see connected to a piece of news and then it

:13:21.:13:24.

would send a signal back to the news distribution platform like Google

:13:25.:13:27.

or Twitter so that they can identify quality news out of the fake news

:13:28.:13:31.

There are already projects afoot to try to flag up stories on sites

:13:32.:13:48.

known to generate fake news like this plug-in but the ideas

:13:49.:13:50.

here are not about blacklisting sites or producing automated

:13:51.:13:53.

Both would be massive undertakings and would themselves provoke cries

:13:54.:13:56.

This is more about letting news organisations prove to their readers

:13:57.:14:00.

The journals of the Washington Post, I work with some amazing people.

:14:01.:14:05.

We produce really great stuff and they remain really committed

:14:06.:14:08.

The tools that we are building here are just a way for us

:14:09.:14:12.

to communicate that we are putting in the effort, where our stuff

:14:13.:14:16.

is coming from, who we are talking to.

:14:17.:14:18.

We are trying to create something that would easily allow audiences

:14:19.:14:29.

to verify for themselves what sources we have used.

:14:30.:14:32.

You'd be able to click and see, who did we talk to, and you'd

:14:33.:14:36.

Readers want to feel like journals are being held accountable to them

:14:37.:14:42.

Other ideas here include ways to fight information bias

:14:43.:14:49.

by surfacing articles that support the opposite side of an argument

:14:50.:14:52.

or to look at the likelihood of truth by finding similar articles

:14:53.:14:55.

Many think the reputation of the journalists themselves plays

:14:56.:14:58.

a big part in the trustworthiness of reporting.

:14:59.:15:00.

Italian newspaper La Stampa is suggesting a system

:15:01.:15:02.

where an author is assigned a unique identifier that shows their piece

:15:03.:15:06.

The best ideas won a small prize at the end of the day but it's

:15:07.:15:17.

The hope is that this part-Google funded initiative may lead

:15:18.:15:22.

to a system that helps news outlet stories rank high up on search

:15:23.:15:25.

That said, I can't help wondering if that is actually something that

:15:26.:15:38.

platforms like Google and Facebook really want.

:15:39.:15:40.

Do you think they care what it is they serve to us or do

:15:41.:15:43.

you think, really, as long as you click on it, that's

:15:44.:15:47.

How do we know the motivations of any company?

:15:48.:15:51.

These companies make their money through people clicking

:15:52.:15:54.

on the adverts, so do you think any of the large companies care

:15:55.:15:57.

Based on my conversations with them, I think they do.

:15:58.:16:07.

The argument would be if they start being perceived as not caring

:16:08.:16:10.

about the information that's sent out there,

:16:11.:16:12.

Yes, it may be that the truth will out, not because of a desire

:16:13.:16:25.

for the facts, but because everyone, readers and news aggregators,

:16:26.:16:28.

More and more people are shopping online but still at this time

:16:29.:16:48.

of year, the high street seems pretty chaotic and the retailers

:16:49.:16:51.

So they are trying to create some more engaging experiences.

:16:52.:17:01.

But do they help us or are they just a distraction?

:17:02.:17:05.

Here in London's Covent Garden, 140 shops and restaurants are taking

:17:06.:17:08.

part in creating one huge augmented reality experience.

:17:09.:17:13.

With the help of AR app Blipper, things come to life.

:17:14.:17:21.

It may not have created the personalised shopping experience

:17:22.:17:24.

I dreamt up, but there were some promotional offers presented

:17:25.:17:27.

as virtual Christmas presents almost around the tree.

:17:28.:17:32.

A reindeer hunt and a giant reindeer you can take a selfie with,

:17:33.:17:35.

I have to say, it wasn't quite as cutting-edge as I'd hoped

:17:36.:17:41.

but I suppose it's a bit of light-hearted fun.

:17:42.:17:45.

Rather more purposefully, the signs in windows can be scanned

:17:46.:17:48.

using image recognition, taking you to online content,

:17:49.:17:50.

partly the sort of stuff you'd be able to look at from your sofa

:17:51.:17:54.

Then came our trip to London's Westfield where augmented

:17:55.:18:01.

We've seen technology like this before but now it's actually

:18:02.:18:09.

on the shop floor here at Charlotte Tilbury.

:18:10.:18:12.

This is what's known as the Magic Mirror and this

:18:13.:18:14.

You choose a lipstick and in real-time, you will see

:18:15.:18:19.

Bright red lips, although it doesn't seem to have any around the edges,

:18:20.:18:28.

I've tried the Rimmel app that does something similar on your phone

:18:29.:18:38.

and you can buy things through it, but here you can do it in the store

:18:39.:18:42.

with assistants all around and a whole shop of products that

:18:43.:18:46.

you can test, smell, and after you see what your face

:18:47.:18:49.

looks like on here, you might want to have a go to check that

:18:50.:18:52.

Meanwhile, here at this eBay event, they are taking things

:18:53.:18:57.

The data on this screen represents what is apparently visitors

:18:58.:19:02.

Using what they call facial coding, the camera looks for reactions

:19:03.:19:06.

which these guys reckon you have when you do online shopping.

:19:07.:19:09.

Nice but don't know who I'd give it to.

:19:10.:19:12.

And I've been told to overact my reactions.

:19:13.:19:27.

Whilst my results bore absolutely no correlation to what I'd liked,

:19:28.:19:32.

maybe they were the ones I contorted my face to the most.

:19:33.:19:35.

Maybe it would have worked better if I'd reacted more naturally,

:19:36.:19:38.

although I struggle to imagine that my face would have shown anything.

:19:39.:19:41.

It left me wondering whether eBay could be developing this if it

:19:42.:19:44.

worked as something more permanent to assess our feelings when online

:19:45.:19:47.

Either way, don't expect me to look too excited about it.

:19:48.:19:57.

Not that I'm sure the tech would have even noticed.

:19:58.:20:02.

Now, in 2015, members of the United Nations adopted a set

:20:03.:20:05.

Number one on that list is to end poverty and to achieve that goal,

:20:06.:20:10.

you first need to work out where poverty exists and how

:20:11.:20:13.

We met up with some scientists at Stamford who have that

:20:14.:20:23.

Marshall Burke is a professor of earth systems science

:20:24.:20:26.

at Stanford University but he spends much of his time in Africa

:20:27.:20:30.

The way this is done is to elicit from the household a listing

:20:31.:20:38.

of everything they've consumed in the last week,

:20:39.:20:40.

So literally everything they've consumed.

:20:41.:20:47.

Every single thing and the value of that item and then you add up

:20:48.:20:51.

all these items for every single person in the household.

:20:52.:20:53.

This can take hours and hours just for one single household.

:20:54.:20:56.

Then you have to do this for thousands of households to get

:20:57.:21:00.

It's painstaking work but Burke has teamed with computer science

:21:01.:21:06.

Using machine learning to predict poverty data

:21:07.:21:12.

But to find out whether the people living in those areas are rich

:21:13.:21:18.

or poor, the researchers used a process called transferred

:21:19.:21:20.

learning and this image of the Earth at night.

:21:21.:21:26.

The parts of the world that are lit up are typically the wealthier parts

:21:27.:21:30.

So basically we use the lower resolution night-time images to help

:21:31.:21:34.

us figure out what in the really high resolution daytime images

:21:35.:21:37.

we should should be using and then we use that to predict poverty

:21:38.:21:40.

Between 300 and 400,000 images were used to train the algorithm.

:21:41.:21:52.

The algorithm will figure out what's important,

:21:53.:21:53.

So some of the things it finds are things that

:21:54.:21:58.

you or I would recognise, things like roads,

:21:59.:22:00.

Based on those features, the algorithm can predict

:22:01.:22:03.

Things like refrigerators, cars, the sum of all those assets.

:22:04.:22:10.

It can also be used to predict incomes.

:22:11.:22:13.

These poverty maps show the team's findings.

:22:14.:22:15.

In areas marked red, people spend as little

:22:16.:22:17.

In green regions like Uganda's capital, Kampala, they spent

:22:18.:22:21.

We are providing a very cheap and scalable alternatives

:22:22.:22:30.

to traditional means of data collection.

:22:31.:22:31.

Traditionally, you have to send people out into the field

:22:32.:22:34.

with clipboards, the surveys aren't always accurate,

:22:35.:22:36.

Like for example lots of governments where they are underperforming,

:22:37.:22:40.

All we need to make our predictions are satellite images.

:22:41.:22:45.

But can you draw conclusions about the economic well-being

:22:46.:22:48.

of communities in Africa when you're thousands of miles away,

:22:49.:22:50.

sitting at a laptop in an office at Stamford?

:22:51.:22:53.

We actually have really good survey information in a few locations.

:22:54.:22:59.

We can use the satellite imagery to make a prediction about poverty

:23:00.:23:02.

and then we can compare that to what the survey says was actually

:23:03.:23:05.

So we used a couple of the really good surveys we had to validate

:23:06.:23:11.

To be a truly useful tool though, the algorithm needs an upgrade.

:23:12.:23:20.

We would also like to use historical imagery so maybe we can figure out

:23:21.:23:24.

how poverty dynamics work overtime and even give us the chance

:23:25.:23:27.

of predicting what's going to happen in the future.

:23:28.:23:31.

If you can pinpoint poverty on a map, aid could be distributed

:23:32.:23:35.

more evenly, policies could be more effective.

:23:36.:23:37.

A picture may be worth a thousand words but combining that picture

:23:38.:23:45.

with artificial intelligence could make a world of difference.

:23:46.:23:52.

That was Sumi and that's it for this week.

:23:53.:23:54.

You can follow us on Twitter @BBCClick for backstage fun

:23:55.:23:57.

and photos and extra technology news throughout the week.

:23:58.:24:00.

Thanks for watching and we'll see you soon.

:24:01.:24:31.

Well, it's still very mild and murky out there.

:24:32.:24:34.

Download Subtitles

SRT

ASS