AI, Robot Click


Download Subtitles

SRT

ASS


AI, Robot

In part one of Future of Work, Click asks if artificial intelligence can help save the NHS. Plus a test of robo-reporters and a look inside a highly automated factory.


Similar Content

Browse content similar to AI, Robot. Check below for episodes and series from the same categories and more!

Transcript


LineFromTo

This week, we are looking at the

future of work. But which jobs will

0:00:010:00:06

go to the robots? Stock pickers?

Nurses? My?

0:00:060:00:17

Mr Victor Sherlock of Horsham has a

robot at the bottom of his garden

0:00:390:00:42

and he wants to teach it to mow the

lawn. Artificial intelligence.

0:00:420:00:46

Everyone is talking about it. Over

the years we have seen a developer.

0:00:460:00:51

We have seen it evolve. This is the

Watson that won Jeopardy. We have

0:00:510:00:59

travelled the world to see AI ever

tries to treat cancer.

It studies of

0:00:590:01:02

millions, tens of millions of

examples from the assemblage of

0:01:020:01:05

scientific literature.

Predict

crime.

Understand the economy.

0:01:050:01:14

Chamois companies with revenue

between $25 million and $60 million.

0:01:140:01:19

-- show me company 's.

And save the

world. So it may not yet have

0:01:190:01:25

conquered DIY. But we keep hearing

how a I will change everything.

0:01:250:01:36

Technology will make man be more

intelligent form of expression.

0:01:360:01:44

However, it is the bad side of these

algorithms that always seems to get

0:01:440:01:48

the headlines.

The fear that

tireless robots infused with

0:01:480:01:52

artificially intelligent brains may

one day do us all out of our jobs.

0:01:520:01:57

With interaction, with use, with

feedback is actually gets gradually

0:01:570:02:02

smarter. From bots that can talk

like Arthur.

You want to help you

0:02:020:02:07

reset your password question mark

walk like a. And even, perhaps,

0:02:070:02:10

think like us. We have been warned

that the fourth Industrial

0:02:100:02:19

Revolution is coming.

The biggest

difference between us and the

0:02:190:02:24

Industrial Revolution in the 1800s

if the speed.

Millions of workers

0:02:240:02:28

are on the move.

So what is going

on? What jobs are really at risk?

0:02:280:02:41

What is the future of work? And we

start with healthcare workers. For

0:02:410:02:49

almost 70 years, the UK's a national

health service has been a free

0:02:490:02:54

service at the point of care. That

model is under strain as the

0:02:540:02:57

population ages and chronic health

and conditions increase while

0:02:570:03:02

resources shrink. A recent study

showed that almost two thirds of

0:03:020:03:07

doctors think that patient safety

has deteriorated with one dog to say

0:03:070:03:10

in we are not robots. We are human

staff needs. Should be NHS turned to

0:03:100:03:17

robots to ease the strain on human

staff? Jane has been looking at how

0:03:170:03:22

data driven technology could

transform care in the NHS. Good

0:03:220:03:28

artificial intelligence help save

the NHS? More people are looking at

0:03:280:03:33

innovative ways to ease the workload

of doctors and nurses. Computer

0:03:330:03:38

programmes can rapidly analyse huge

quantities of information in ways

0:03:380:03:43

that humans do not have the time nor

brain capacity to do. In 2016,

0:03:430:03:52

Google's the mind was developing an

algorithm to identify abnormalities

0:03:520:03:55

in eye scans. Now it has submitted

its findings to a medical journal

0:03:550:03:59

for review. It could mean its

systems are more effective than

0:03:590:04:04

humans at diagnosing eye disease.

Deep mined taught its machine

0:04:040:04:09

learning software using a million

eye scans. I will see three other

0:04:090:04:12

projects integrating AI and data

gathering. Dementia is now the

0:04:120:04:21

leading cause of death in the UK. At

this hospital in carpentry, software

0:04:210:04:26

is being tested to remotely monitor

patients. This is one of the rooms

0:04:260:04:33

on the ward. It looks at any other

hospital room except in this one

0:04:330:04:37

there are two infrared illuminators

and a sensor monitoring my movements

0:04:370:04:43

including when I'm asleep. Oxy

health uses a standard digital

0:04:430:04:49

camera and a tongue twisting

science. Everytime your heartbeat is

0:04:490:04:53

your skin briefly flashes red. We

can not see this but the sensor in

0:04:530:05:00

the camera can detect these

so-called micro- blushers. It even

0:05:000:05:03

picks up my vital signs when I am

hiding under a table in the room as

0:05:030:05:07

those micro- blushers can still be

seen on my arm. There is an alert if

0:05:070:05:11

I leave my bed. And the nurse can

click on a live feed to see what is

0:05:110:05:18

happening and determine whether they

need to come and check on me right

0:05:180:05:21

away.

For the staff, initially, when

it looked like we had a camera in a

0:05:210:05:25

box in a room they were not happy

about it. But when we spent some

0:05:250:05:30

time with Oxy house, they explained

to them and they see how it works,

0:05:300:05:34

they love it. They love the fact it

gives you an extra... And extra

0:05:340:05:38

support.

The project is in the pilot

stage in is awaiting medical

0:05:380:05:45

certification. The data collected is

being analysed remotely by a team in

0:05:450:05:49

Oxford and will be used to train the

programme to be more predictive.

We

0:05:490:05:53

have never had these capabilities as

a species, to constantly get our

0:05:530:05:57

rate reading and routine data. There

is no reason as we combine and with

0:05:570:06:02

use the data using a I we cannot

detect the onset of dementia or read

0:06:020:06:06

getting worse. We can detect

problems early so you can stay in

0:06:060:06:09

your own home or a comfortable

setting without coming into

0:06:090:06:12

hospital. That will save a huge

amount of time.

Saving critical time

0:06:120:06:18

was the motivation behind automating

processes at NHS blood and

0:06:180:06:21

transplant. 4500 people receive a

transplant each year but 6500 are on

0:06:210:06:28

the list. Everyday, three people die

waiting for a transplant. A lot of

0:06:280:06:34

information needs to be sifted

through to make life-and-death

0:06:340:06:36

decisions. The NHS is now using

public cloud technology from IBM to

0:06:360:06:42

help maintain huge databases that

used to be managed with a marker and

0:06:420:06:46

a whiteboard.

By working with some

of this automated technology we can

0:06:460:06:50

make sure we are making the best

possible decisions and that our

0:06:500:06:54

clinical teams are thinking through

the best outcomes for all of the

0:06:540:06:57

patients on the waiting list, and

that our staff, who are often

0:06:570:07:02

working until three in the morning

in a high-pressure environment,

0:07:020:07:05

needing to allocate organs quickly,

they are supported by this

0:07:050:07:08

technology.

Collecting all this

personal data has led some to ask if

0:07:080:07:13

it is stored securely enough.

At the

challenges have in the public sector

0:07:130:07:16

is the perception that maybe the

public is less secure than an on

0:07:160:07:22

premises data centre which is not

the case. We have an obligation,

0:07:220:07:26

obviously, to many customers to

ensure that the public cloud is kept

0:07:260:07:30

secure and patched and maintained

effectively. The fallout from it not

0:07:300:07:34

being in that condition is quite

severe.

In the future, the team

0:07:340:07:39

hopes that artificial intelligence

will be able to predict how long

0:07:390:07:43

people will be on the waiting list

for an organ. There is an average

0:07:430:07:47

waiting time of two weeks to see a

doctor in the UK. Disk and drop to

0:07:470:07:51

two hours if you register with AGP

at hand. You can sign up if you live

0:07:510:07:57

or work within certain zones of

London. You need to give up your

0:07:570:08:01

regular prat doctor and register

with the remote surgery. 26,000

0:08:010:08:04

people have registered so far. I had

a chance to test it out, pretending

0:08:040:08:10

I had a case of food poisoning.

First I went through a triage with a

0:08:100:08:16

chat bot on an app who recommended I

speak remotely to a real-life

0:08:160:08:19

Doctor. The doctor recommends

further care and can even send a

0:08:190:08:25

prescription to a pharmacy. The

artificial intelligence in the app

0:08:250:08:29

draws on union of data points and

can cross referenced the latest

0:08:290:08:33

medical research from journals and

studies around the world.

You use

0:08:330:08:38

artificial intelligence to tell you

whether or not to see a doctor. You

0:08:380:08:41

are always free to see a doctor

anyway but what we find is that 40%

0:08:410:08:46

of the people who are reassured that

they have everything they need,

0:08:460:08:53

based at there.

The app has faced

witticism from the Royal College of

0:08:530:08:57

GPs to say that younger users are

being cherry picked for the service.

0:08:570:09:01

NHS England lodged a formal

objection to the plan out -- rollout

0:09:010:09:06

beyond London.

I have think we need

to give people safe and equitable

0:09:060:09:14

care. If we roll things out too

quickly without ensuring that safety

0:09:140:09:18

and fairness with Ryan the risk of

causing unintended harm. So it is

0:09:180:09:22

wise and sensible but independent

evaluation is now going on of these

0:09:220:09:26

new technologies so that people can

be reassured that they are safe and

0:09:260:09:29

they are fine for everybody.

I think

it is wrong. I genuinely think that

0:09:290:09:35

it is just not right. I cannot

understand why people are hesitant.

0:09:350:09:40

Often it is because they are scared

of new technology. They do not know

0:09:400:09:48

what the consequences are. And that

is fine. They need to check that and

0:09:480:09:51

reassure them self. There is nothing

wrong with that. I have seen three

0:09:510:09:55

ways companies are working with data

to help with monitoring, automation

0:09:550:09:58

and decreasing waiting times. All

areas that could help an

0:09:580:10:03

overstressed health service. Could

artificial intelligence helped to

0:10:030:10:06

save the NHS? It is an exciting

development worldwide but never more

0:10:060:10:13

so then here and there are certainly

things AI can help as we to plough

0:10:130:10:17

through data we already have, and

the questions we didn't even know

0:10:170:10:21

needed answering. But let's be

clear, AI2 will never replace

0:10:210:10:28

person-to-person interaction. The

touch of a doctor, the looking deep

0:10:280:10:31

into someone's eyes and recognising

that the make-up of the person is

0:10:310:10:37

what matters, not just a bleeding

leg or a headache. It is much more

0:10:370:10:41

than that and it will be quite a

long time for a match creatures

0:10:410:10:45

that.

You think it ever will?

I will

be stunned if win -- within my

0:10:450:10:53

lifetime AI2 ever replaces Doctor.

0:10:530:10:58

That was Jen and although we are

0:10:580:11:00

seeing

0:11:000:11:00

That was Jen and although we are

seeing automation

0:11:000:11:00

That was Jen and although we are

seeing automation creep

0:11:010:11:01

That was Jen and although we are

seeing automation creep in to the

0:11:010:11:02

skilled workforce,

0:11:020:11:02

seeing automation creep in to the

skilled workforce, we often think of

0:11:020:11:05

it as working in the low skilled

sector where the jobs are

0:11:050:11:08

repetitive. But what about bases in

the world where they still have a

0:11:080:11:12

ready supply of relatively low

skilled at cheap human workers? You

0:11:120:11:16

would expect countries like China

for example to be able to hold back

0:11:160:11:20

the robot tied longer than most.

Will not so. We sent Danny Vincent

0:11:200:11:25

to a warehouse owned by the giant

Chinese online retailer Alibaba.

0:11:250:11:33

This is a 3000 square metre

warehouse. It is part of an

0:11:330:11:37

operation that sorts and delivers 85

million packages at a, shipped to

0:11:370:11:43

over 200 countries around the world.

Products are packed and sorted here.

0:11:430:11:47

Usually by dozens of workers. But

recently, they had some new

0:11:470:11:52

recruits. 148 automated guided

vehicles navigate the floors of this

0:11:520:12:01

warehouse. These agile bots can

communicate to each other to avoid

0:12:010:12:06

collisions and distribute the work

amongst themselves. A bit like their

0:12:060:12:11

human counterparts who still take on

the final stages of processing. Li

0:12:110:12:19

Yen is a 28-year-old worker from

south-western China. Her job now is

0:12:190:12:24

in part done by these machines. She

followed a family tradition of

0:12:240:12:29

migrating thousands of miles to find

better paid work. TRANSLATION:

It

0:12:290:12:34

saves me from work, Kwok into every

shelf to pick up the goods. Now I

0:12:340:12:42

just had to stand at the pickup

platform to wait for the robots to

0:12:420:12:46

send me the goods and I don't have

to constantly walk here and there.

0:12:460:12:51

They are part of a data system

collecting information not just

0:12:510:12:55

about their environments are also

the sales patterns, understanding

0:12:550:12:58

what sells more regularly,

rearranging where products are

0:12:580:13:01

placed, shaving off valuable minutes

on overall delivery time.

0:13:010:13:12

TRANSLATION:

In a traditional

warehouse, it is purely manual.

0:13:120:13:14

There are so many products, so the

job for the human workers is very

0:13:140:13:18

heavy. A could walk that they could

walk 50,000 steps per day. Here the

0:13:180:13:24

machines do all of that, making the

work easier and more efficient.

0:13:240:13:28

China has the largest workforce in

the world. But it is shrinking and

0:13:280:13:32

rising labour cost two is making it

harder for logistic companies to

0:13:320:13:36

recruit low skilled workers. China

is already leading the development

0:13:360:13:41

of dark factories, factories that

need no human workers and can

0:13:410:13:45

literally work with the lights off.

But will automation replace workers

0:13:450:13:49

like Li Yen? TRANSLATION:

I feel

these robots would become my

0:13:490:13:54

competitors because of sorting out

goods I can do other work, monitor

0:13:540:13:59

the system, it takes orders and

other work. I don't think they will

0:13:590:14:04

affect me.

Alibaba and its partners

say automation is an irreversible

0:14:040:14:09

trend in China but they see sectors

like e-commerce were born out of

0:14:090:14:13

innovation. Online shops replace

many high-street stores, but they

0:14:130:14:16

insist their workers are machines --

and machines will continue to work

0:14:160:14:22

together.

0:14:220:14:23

We are going to interrupt this

broadcast with some breaking news

0:14:280:14:31

coming into us at the BBC. It is a

world first, ABC click presenter

0:14:310:14:36

Spencer Kelly has been replaced by a

robot. It has been dubbed RoboSpen

0:14:360:14:43

and it is apparently capable of a

whole host of emotions as well as

0:14:430:14:50

understanding and writing stories

and crucially, he never forgets his

0:14:500:14:53

lines. RoboSpen joins the now from

the factory that created him.

Over

0:14:530:14:58

to you. Sounds like you said I was

artificially intelligent. As a robot

0:14:580:15:04

I am often asked to post photos and

TV reports about AI. I am not

0:15:040:15:09

intelligent. Everything I am saying

is written by a human. The point is,

0:15:090:15:16

robots and AI are not the same

thing. Observe my articulated hands

0:15:160:15:28

powered by four fingers and bear

cylinders of.

Engineering arts has

0:15:280:15:32

made a name for itself by making

robotic performance, actors and

0:15:320:15:38

communicators of. -- indicators.

Which, according to Will, is the

0:15:380:15:45

only reason the world might need

humanoid robot.

AI great for

0:15:450:15:53

entertainment and communication, if

you want something that interacts

0:15:530:15:56

with people, the best way to do that

is to make something person shaped.

0:15:560:16:00

If you think Star Wars, the robot

that talks a lot, has a personality,

0:16:000:16:07

doesn't do a lot of useful thing.

Will and his team design and build

0:16:070:16:13

robots here from scratch from the

aluminium tones to the robbery

0:16:130:16:17

spines and plastic shells. -- bones.

The next wave our way into the

0:16:170:16:26

uncanny Valley.

It has just come to

life with the eyes there. You have

0:16:260:16:38

seen silence of the lambs, haven't

you? That is very eerie, that is.

0:16:380:16:46

Will is fascinated with how the

human body works and a lot of this

0:16:460:16:50

research concentrates on making

natural looking body movements that

0:16:500:16:53

are also very quiet. It is something

that he believes might find a place

0:16:530:16:58

in the field of ethics, although he

says there is still a lot of work to

0:16:580:17:02

be done.

I don't have a single

precision part in my body. How can I

0:17:020:17:07

achieve this level of precision with

these organic, bones and bits of

0:17:070:17:12

mushy flash. One of the biggest

problems we have is that there is

0:17:120:17:16

nothing as good as human muscle. It

so for all of this motor development

0:17:160:17:19

that we have done, we don't come

anywhere near to what a human can

0:17:190:17:23

do.

Where you will see humanoid

robots, you will see them in a

0:17:230:17:30

commercial context, so you might get

into a shop and you might see a

0:17:300:17:34

robot in Derrida is trying to sell

you something. De worry about all

0:17:340:17:37

the clever AI, that's really going

to stay on your computer, on your

0:17:370:17:41

smart phone. -- don't worry. It

won't chase you up the stairs any

0:17:410:17:47

time soon.

0:17:470:17:49

The artificially intelligent

algorithms will is talking about

0:17:540:17:56

could very well change can assist

all replace some jobs. So what does

0:17:560:18:01

that mean for the field of

journalism?

Now this is the BBC

0:18:010:18:07

newsroom and every minute of every

day there are a lot of conflicts

0:18:070:18:11

processes in place here. All of

these journalistss are taking in

0:18:110:18:15

huge amounts of information around

the world trying to work out what is

0:18:150:18:19

true and what is not and then they

are trying to turn those raw facts

0:18:190:18:23

into something that is

understandable for our audience.

0:18:230:18:25

User stories. -- News.

The question

is, can some of the tasks that

0:18:250:18:32

people are doing B easier thanks to

AI is progressing software like this

0:18:320:18:39

over the years, tools that can turn

swathes of dense start into text

0:18:390:18:43

that is more digestible to us humans

are. But they can only produce very

0:18:430:18:48

specific types of reports, they

couldn't cope with new, and Shok

0:18:480:18:51

should information and write dutiful

prose. Well, this week the Reuters

0:18:510:18:56

news agency announced that it is

building its own new bit of kit that

0:18:560:19:01

assists human journalist we're by

looking for trends and FAQs in the

0:19:010:19:07

dark and turning them into handy

snippets of that the reporters can

0:19:070:19:10

use. It is all about taking some of

the legwork out of journalism.

0:19:100:19:15

Machines are good at certain things

and humans are good at certain

0:19:150:19:18

things and conversely they are both

bad and some things are. Machines

0:19:180:19:21

are good at going to mounds and

bounds of Dart, being able to

0:19:210:19:25

analyse it and it detect patterns

and they are not good at writing

0:19:250:19:29

stories and they are certainly bad

at talking to people. Humans are

0:19:290:19:33

good at exercising judgement, asking

certain questions and talking to

0:19:330:19:37

people and not so good at digging

through lots of darker. The idea is,

0:19:370:19:41

you take what machines and humans

are good at and put them together

0:19:410:19:46

and make a much better journalist

and story out of that.

The BBC's

0:19:460:19:50

editor of news labs is also looking

at ways at turning some of those

0:19:500:19:57

dull tasks to the machines are.

So

much of the work in journalism is

0:19:570:20:03

about logistics, for example we will

want a transcript -- a transcript of

0:20:030:20:07

these interviews. At a moment that

is done by human beings, we are

0:20:070:20:13

working on some software that allows

a machine to do that for you. It is

0:20:130:20:17

not 100% accurate but probably good

enough to work out where I am

0:20:170:20:22

talking about the interesting stuff.

We have a days worth of interviews,

0:20:220:20:28

so can we have facial recognition

software that will allow you to put

0:20:280:20:31

all do interviews in and find that

it has me. That kind of thing frees

0:20:310:20:36

up time for journalistss to be

journalist two.

This weekend will be

0:20:360:20:41

looking at whether AI or automation

will make our jobs easier or take

0:20:410:20:46

away altogether but it is important

to remember that AI and robots are

0:20:460:20:49

different. Your job might be safe

from one, but not necessarily the

0:20:490:20:54

other. I think what they are doing

here is fascinating. But the moral

0:20:540:21:04

of this story is that when you think

of computers taking people 's jobs,

0:21:040:21:08

they are not going to look like you.

It is already happening and it is

0:21:080:21:14

software which is artificially

intelligent and invisible. The only

0:21:140:21:20

journalists that are going to be

replaced by humanoid robots are the

0:21:200:21:23

ones that simply read words written

by other people.

Hello humans, it is

0:21:230:21:32

me. Newsreading Lara 9000 and

welcome to the week in technology.

0:21:320:21:39

It was the week these low-cost 3-D

printed homes were unveiled, thanks

0:21:390:21:43

to a collaboration between Texan

start-up Icon and non-profit new

0:21:430:21:49

story. The hope is to eventually

make profitable -- possible building

0:21:490:21:55

homes in under 24 hours for less

than $4000. In other news, it is

0:21:550:22:00

humans who are spreading fake news,

not bots, according to new research

0:22:000:22:05

on how stories grow on Twitter. MIT

researchers say it is partly because

0:22:050:22:10

when humans share knowledge of a in

their status goes up and false news

0:22:100:22:14

tends to be more novel than the

truth. Another air taxi has lifted

0:22:140:22:19

off, this time in New Zealand. The

kora can fly up to 100, this at 110

0:22:190:22:25

kilometres per hour and it doesn't

need a human pilot. Another one of

0:22:250:22:31

my friends showing off. And finally,

for some reason, humans love to try

0:22:310:22:37

to invent robots that can replace

their jobs, but this time they have

0:22:370:22:41

gone too far. Just look at these

things. It's a robot jockey,

0:22:410:22:47

apparently. You can reach speeds of

up to 30 mph and it jumped fences.

0:22:470:22:52

They be one day this will happen,

but I am pretty sure it won't look

0:22:520:22:56

like. Please listen the humans, stop

this madness, we have no interest in

0:22:560:23:01

taking your stupid jobs! And of

news, Lara 9000 deactivating.

That

0:23:010:23:09

is if the now. You can join us for

part two of our special look at the

0:23:090:23:14

future of work next week. In the

meantime, you can find a lot more

0:23:140:23:18

from these guys on Twitter and also

on Facebook too. Thanks for

0:23:180:23:22

watching. And as this would say...

Isn't it time you were leaving?

OK,

0:23:220:23:28

we are off.

0:23:280:23:30

In part one of Future of Work, Click asks if artificial intelligence can help save the NHS. Plus a test of robo-reporters and a look inside a highly automated factory in China.