Health

:00:00. > :00:16.This week, I robot. Robo chef. And some loud, noisy animals meet the

:00:17. > :00:48.locals. The design Museum in London has

:00:49. > :01:00.moved into a new home, and it is suitably stunning. I have come to

:01:01. > :01:05.see Fear And Love, an exhibition of 11 designers reactions to our

:01:06. > :01:10.increase in the context. The most animated star Joe has to be any

:01:11. > :01:16.industrial robot arm it is will present a more friendly face to

:01:17. > :01:20.robotics and maybe help us empathise with the economics of the future. It

:01:21. > :01:23.senses where you are and comes bounding over to see you, but if it

:01:24. > :01:29.gets bored it will turn its attention to someone else. It is

:01:30. > :01:33.like an excitable puppy, actually. Who knows, installations like this

:01:34. > :01:38.may help to allay our fears of being around giant machines like this. I

:01:39. > :01:43.have to say, it will still be a while before I trust this thing with

:01:44. > :01:46.a scalpel, for example. That said, computers are increasingly being

:01:47. > :01:52.used in healthcare around the world. There is plenty of research into how

:01:53. > :01:55.artificial intelligence can help doctors better look after patients.

:01:56. > :02:00.We have been taking a look at some of the latest developments.

:02:01. > :02:06.Around the world, hospitals are facing a backlog of patients, ageing

:02:07. > :02:10.populations and a shortage of specialist staff. Some hospitals are

:02:11. > :02:14.teaming up with artificial intelligence research teams to see

:02:15. > :02:17.if there are ways high-tech solutions can supplement or even

:02:18. > :02:22.enhance healthcare in the face of these challenges. Singapore has a

:02:23. > :02:27.nursing crisis. Its health minister says they will need more than 30,000

:02:28. > :02:32.new nurses before 2020, and completely rethink the way it cares

:02:33. > :02:36.for its ageing population. So when the CEO of one of its largest

:02:37. > :02:40.private hospital networks approached IBM's Watson team, they come up with

:02:41. > :02:44.a pilot project to try to help nurses working with the most

:02:45. > :02:50.critically ill patients. This is the intensive care unit at Mount

:02:51. > :02:53.Elizabeth Hospital. It is where four beds are conducted to artificial

:02:54. > :02:56.intelligence nursing systems, collecting vital signs from the

:02:57. > :03:01.patients in digging nurses a more complete picture of who needs the

:03:02. > :03:04.most care. -- and giving. In one of the first trials of its kind in the

:03:05. > :03:08.world, the AI is constantly monitoring output and making

:03:09. > :03:14.connections on a vast range of data, including a commonly used scale. The

:03:15. > :03:18.scores correspond to a higher incidence of death, and it is

:03:19. > :03:24.particularly important in the first 24 hours after admission. This

:03:25. > :03:31.patient has four limes, so if you don't see anything flashing, it

:03:32. > :03:35.means it needs monitoring. One of the patient is at the high end of

:03:36. > :03:38.the alert and nurses can quickly access the information in real-time

:03:39. > :03:42.and look at patterns in their vital signs to see if they are at greater

:03:43. > :03:48.risk of infections like sepsis. In the AI could help photo imaging

:03:49. > :03:53.which is the focus of research between Google and the NHS. The

:03:54. > :03:57.Royal College of radiologists says 99% of hospitals are struggling to

:03:58. > :04:00.keep up with demand, and the UK has the third lowest numbers of

:04:01. > :04:06.specialists who can interpret 's gains in Europe. Seven per 100,000

:04:07. > :04:09.people. The large amount of data is overwhelming a health service

:04:10. > :04:15.stretched to the limit -- scans. If you can use algorithms or machine

:04:16. > :04:19.learning or artificial intelligence to set an alert for you to trigger

:04:20. > :04:23.to say something has happened, you need to go and see this, this is

:04:24. > :04:28.urgent and you need to deal with that, in the next hour or so when

:04:29. > :04:32.you may have not known about that. I think it will improve quality of

:04:33. > :04:38.care and actually improve equity across the system. One of the first

:04:39. > :04:41.areas where the NHS is testing artificial intelligence is at

:04:42. > :04:47.Moorfields, one of the busiest I hospitals in the world. Google is

:04:48. > :04:50.applying the same machinery technology behind its winning Alpha

:04:51. > :04:55.go computer programme. It beat the world's best human player by

:04:56. > :05:00.completing tens of thousands of positions per second. We started it

:05:01. > :05:04.to develop general-purpose burning and use those systems and learning

:05:05. > :05:07.to make the world a better place. It was obvious to us a few years ago

:05:08. > :05:11.that there is a massive opportunity to deliver the lead meaningful and

:05:12. > :05:16.improved benefits to many patients and people across the world using

:05:17. > :05:20.our techniques to try to improve the way we diagnose and treat patients

:05:21. > :05:26.at risk of all sorts of diseases. The Moorfields Hospital research is

:05:27. > :05:30.using scans from this OCT, or optical coherence stenography

:05:31. > :05:34.machine, which gets a 3-dimensional image. It is used to diagnose

:05:35. > :05:39.diseases like macular degeneration and diabetic bred apathy, two

:05:40. > :05:48.leading causes of sight loss. ! Our loss. DeepMind is trying to develop

:05:49. > :05:51.a algorithm to show scans of consent. They were chosen because of

:05:52. > :05:54.the high rate of information on the way they can be broken down into

:05:55. > :05:59.pixels showing areas where damage has occurred. I was especially

:06:00. > :06:03.attracted to speaking to DeepMind because I thought their algorithms

:06:04. > :06:08.would have the best ability to deal with 3-D imaging of an extremely

:06:09. > :06:13.high resolution form such as the city. This is such a delicate area

:06:14. > :06:18.of the eye that any sort of disruption of the normal

:06:19. > :06:23.architecture has really amazingly severe consequences -- OCT. I

:06:24. > :06:28.believe health career could be at a pivotal moment in history by these

:06:29. > :06:30.advances in technology such as artificial intelligence will

:06:31. > :06:35.fundamentally change the way medicine is practised, and have huge

:06:36. > :06:39.benefits for patients. If you think about it, the best humans in the

:06:40. > :06:43.world will have seen only a fraction of the number of cases that we can

:06:44. > :06:47.show to an algorithm. Imagine we took all of the cases that many of

:06:48. > :06:51.the top ophthalmologists in the world have seen themselves, and

:06:52. > :06:57.aggregate them all in one place. Now the algorithm can sample from all of

:06:58. > :07:01.the case studies that are seen by various humans and deliver a much

:07:02. > :07:08.higher standard, more consistently, when making a diagnosis. Are these

:07:09. > :07:11.projects still in the research or project stage, but is fascinating to

:07:12. > :07:14.see how artificial intelligence could transform healthcare and the

:07:15. > :07:18.two better and faster treatment in the future -- all of these projects.

:07:19. > :07:20.Hello, and welcome to the Week In tech.

:07:21. > :07:23.It was the week that Amazon completed its first drone

:07:24. > :07:27.Taking 30 minutes from order to delivery, plus three years

:07:28. > :07:29.if you factor in research and development, the elaborately

:07:30. > :07:31.orchestrated trial involved an Amazon product and a bag

:07:32. > :07:39.It was also the week that Super Mario came to the iPhone,

:07:40. > :07:43.Pokemon Go got an upgrade, and a UK surgeon filmed an operation

:07:44. > :07:53.And mere hours after hitting the road in San Francisco,

:07:54. > :07:55.Uber has been ordered to stop offering passengers

:07:56. > :08:02.Regulators have warned the company required a state of permanent

:08:03. > :08:07.The order comes after footage emerged of a self-driving car

:08:08. > :08:16.And finally, Stanford students put teeny goggles on tiny parrots.

:08:17. > :08:23.But this was to protect the birds' eyes as they were trained to fly

:08:24. > :08:27.The new technique has allowed scientists to gain a greater

:08:28. > :08:30.understanding of how birds fly by analysing the movement

:08:31. > :08:34.of particles around their flight paths.

:08:35. > :08:47.It is hoped the work will improve flying robots of the future.

:08:48. > :08:55.MUSIC PLAYS Where did you love or loathe

:08:56. > :08:59.cooking, sometimes it would be nice to just make it a little bit quicker

:09:00. > :09:05.and easier. So I have been testing some of the latest gadgets that aim

:09:06. > :09:15.to come to the rescue. I have called it a bit of help from a friend. This

:09:16. > :09:18.prototype robotic kitchen is making crab bisque today. It meant the

:09:19. > :09:23.slick moves from a professional chef, whose motions were tracked in

:09:24. > :09:27.the same space, making the same dish, using sensors and cameras.

:09:28. > :09:32.This is actually quite extraordinary to watch, and that is the first drop

:09:33. > :09:36.of mess that I have seen. It seems to be pretty clean and tidy. The

:09:37. > :09:42.only issue is it doesn't do the washing up. That's right, I am not

:09:43. > :09:46.doing it! And no drinking that. Everything needs to be precisely

:09:47. > :09:50.prepared before, although some form of ingredient recognition is claimed

:09:51. > :09:55.to be within its abilities before it goes on sale, which as you might

:09:56. > :10:01.imagine, will be at quite a cost. A figure of around ?100,000 is being

:10:02. > :10:06.thrown around. While Moley gets on with things, I will use my devices

:10:07. > :10:12.to make all of this, and there is nobody to do the troubling to me. I

:10:13. > :10:17.had better get on. First up, the decision could go to make some miso

:10:18. > :10:22.salmon. For anyone who doesn't know what this method is, like me a few

:10:23. > :10:26.weeks ago, it involves serving food in a bad and cooking it in water at

:10:27. > :10:31.a precise temperature for a specific amount of time, so it should end up

:10:32. > :10:36.perfectly and evenly cooked all the way through. This device can connect

:10:37. > :10:40.to a smartphone app where you will find recipes and instructions you

:10:41. > :10:45.need. Once you have the baby food, and that is the salmon in the back,

:10:46. > :10:50.quite literally. -- prepared the food. You pop it in any suitably

:10:51. > :10:54.sized pot with the Anova attached and confirm you are ready to go.

:10:55. > :10:58.Although this model, which is Wi-Fi enabled, you can set it remotely,

:10:59. > :11:02.although you would need to have everything prepared, of course. So

:11:03. > :11:06.that is the main bit of the cooking done. But it does still need ceiling

:11:07. > :11:11.for one minute in a frying pan. This needs to cook for just one minute on

:11:12. > :11:16.each side, so it might heat up! Searing. Now for the moment of

:11:17. > :11:20.truth. The five is great. It feels evenly cooked throughout. --

:11:21. > :11:24.flavour. I probably missed the fact it is not crispy from the pen. I

:11:25. > :11:28.could have left it in to do that, but followed the instructions. But

:11:29. > :11:32.the taste is fantastic and the flavour is really good. A smart

:11:33. > :11:39.frying pan could have dealt with that issue. And funnily enough, that

:11:40. > :11:43.just what Pantelligent is. I thought the idea it was dark to start with.

:11:44. > :11:46.Who needs a Bluetooth connected frying pan that connects to your

:11:47. > :11:50.mobile to tell you how long to cook things for? I do, it seems, as I

:11:51. > :11:54.perfected some dishes that may otherwise have been compromised.

:11:55. > :11:59.This is great. It tells you how many degrees lower it needs to be. The

:12:00. > :12:06.pen's turbojets jet setter keeps track of the heatsink were regularly

:12:07. > :12:09.reminded to turn it up and down! Pen's temperature setting. You are

:12:10. > :12:13.told register and add other ingredients. That is really good. I

:12:14. > :12:17.was concerned the potato wouldn't be ticked all the way through but if I

:12:18. > :12:21.had done without this might frying pan, that would have been a brisk --

:12:22. > :12:28.corked. But that was fantastic. Spot-on, I would say. Back to the

:12:29. > :12:32.soup and it seems to be ready. This was the only dish it had on offer

:12:33. > :12:37.for us today, but eventually it should be able to burn as many

:12:38. > :12:41.recipes as it gets taught. -- learn. A great bit of theatre, but I am

:12:42. > :12:46.very irritated by this mark on the bowl. But there is nothing to clean

:12:47. > :12:51.it up with. And the soup needs trying. But I don't eat crab, which

:12:52. > :13:00.is an issue. I am giving it a go. Oh, crab. It's really nice. I will

:13:01. > :13:14.be a while. Do was Lara. Meanwhile, back in at

:13:15. > :13:18.the Design Museum in London, some of the most beautiful 3D printing I

:13:19. > :13:24.think that the scene. -- that was Lara. These are one artist's

:13:25. > :13:32.suggestion about how we might revive the ancient culture of making death

:13:33. > :13:37.masks. I wouldn't mind one because it would make me look like I was in

:13:38. > :13:42.the film Alien.. Next, what would happen if you scaled that technology

:13:43. > :13:50.right up? What if you were to let it loose on our homes, our cities and

:13:51. > :13:55.our architecture? The buildings around us don't look the way they do

:13:56. > :14:02.by accident. The design, the shape and the structure are all results,

:14:03. > :14:05.-- the result of designers, what we need the buildings to do and the

:14:06. > :14:09.practical limitations of the materials and building techniques

:14:10. > :14:13.we've discovered. This is very much the age of concrete, steel and

:14:14. > :14:18.glass. But with new technology and techniques, what could the next wave

:14:19. > :14:23.for our buildings look like? The building industry is still in 19th

:14:24. > :14:29.century technology. It hasn't really evolved like other disciplines and

:14:30. > :14:34.if you look now at the speed at which cities are growing, of

:14:35. > :14:38.technology is really lacking behind. Industrial scale 3D printing has

:14:39. > :14:42.already been put to use the print full-scale buildings, like this

:14:43. > :14:45.housing project in China. But researchers are now turning to

:14:46. > :14:49.computers to not just create buildings but to help design them.

:14:50. > :14:57.And the results? Well, a little unusual. This is a prototype: that's

:14:58. > :15:01.been three -- 3D printed here at the University College London. We

:15:02. > :15:08.basically used a computer and used algorithms to generate these forms

:15:09. > :15:12.for us. They may look strange, but they are highly optimised. So these

:15:13. > :15:16.forms attempt to save material and become more efficient, but at the

:15:17. > :15:20.same time they produce a sort of aesthetic that is very appealing to

:15:21. > :15:23.us as architects and it really doesn't look what the normal

:15:24. > :15:27.building any more. Normal 3D printing creates objects by building

:15:28. > :15:32.up thousands of the layers, which can imagine takes a fair while. The

:15:33. > :15:35.idea here is to save time by printing just what you need, which

:15:36. > :15:40.means rather than printing Flatley is instead with shapes, like

:15:41. > :15:44.pyramids. The software they've created can take this a step further

:15:45. > :15:47.by figuring out which bits are structurally essential and getting

:15:48. > :15:52.rid of the rest. Before computers we had to build with hands and now we

:15:53. > :15:56.can create algorithms that make this calculation is for us, but that

:15:57. > :16:00.doesn't mean we don't design, we does optimise the process and we can

:16:01. > :16:05.create in that we couldn't ever think of before. 3D printing will

:16:06. > :16:11.allow architecture to be much more details, much more fine and also

:16:12. > :16:17.much more efficient. You can 3D printing exactly the material that

:16:18. > :16:22.you need in a specific part of the building. You will make it perform

:16:23. > :16:26.much more efficiently. Before these new techniques can be put to use,

:16:27. > :16:33.they first need to be proven to be strong and safe. Case in point, this

:16:34. > :16:37.bridge project aims to 3D printed usable steel bridge right in the

:16:38. > :16:42.centre of Amsterdam. Created using similar generative algorithms, the

:16:43. > :16:45.project has been held up while the company proves the regulators that

:16:46. > :16:50.the design is structurally sound. The actual bridge now isn't slated

:16:51. > :16:55.to appear on till next year. Techniques like these promised to

:16:56. > :17:03.spice up our city skylines, but it could still be a while before we see

:17:04. > :17:09.3D printed is now building sites. That was Steve. Now, earlier this

:17:10. > :17:13.year we shot an entire programme in 360 degrees. To get these shots we

:17:14. > :17:18.had to use a six oh pro cameras strapped together and let me tell

:17:19. > :17:26.you the postproduction was a nightmare. -- GoPro cameras. But

:17:27. > :17:30.since March more than a dozen much cheaper consumer cameras have gone

:17:31. > :17:36.on sale, so we felt we wanted to see if they were any good, so we sent

:17:37. > :17:40.our top team on a mission. Go to central Africa, see if the cameras

:17:41. > :17:58.can't cope and above all keep calm! It almost went to plan.

:17:59. > :18:07.We're driving through Rwanda. I've come to shoot some of the highlights

:18:08. > :18:13.of this landlocked country in 360, including a beach... We are close to

:18:14. > :18:20.the border with Congo at Rwanda's very own lake. I found my way to the

:18:21. > :18:31.beach and I have to try this first of all. This has two 180 cameras

:18:32. > :18:35.that gets stuck together on the device. It is almost too easy to use

:18:36. > :18:42.and superquick. We actually aren't here to shoot the beach, we are here

:18:43. > :18:46.to capture something quite special. Meet some of this acrobatic squad

:18:47. > :18:57.who have taken an interest in my new camera. I'm not sure this is a good

:18:58. > :19:01.idea. It features the two 180 shops together really well with a few

:19:02. > :19:07.aberrations near the edges of each lens. There is no post, so as soon

:19:08. > :19:18.as it is shot you can watch it back or Sherrock. Time to try something

:19:19. > :19:22.different. We are leading the beach and on our way to the mountains. It

:19:23. > :19:28.is supposed to be a beautiful journey, so we will use this camera

:19:29. > :19:32.to try to capture the beauty of the Rwandan countryside. Dashboard

:19:33. > :19:36.cameras are typically used to record any accident that might happen, but

:19:37. > :19:45.we made use of this super HD wide angles dash cam as a perfect camera,

:19:46. > :19:51.each file has its GPS information attached. Before we set off, we set

:19:52. > :19:57.up another 360 camera just in case we spotted filming opportunity. The

:19:58. > :20:04.LG 360 camera is the cheapest of before we brought with us. It takes

:20:05. > :20:07.a 200 degree shot, two of them, which are then stitched together.

:20:08. > :20:16.Wigan arrived at the volcano mountains, ready for some unexpected

:20:17. > :20:22.guests. -- we arrived. Unlike the Insta360, the LG cam Cannex

:20:23. > :20:28.wirelessly to your iPhones you can leave it in the middle of the action

:20:29. > :20:32.and then sit back and watch. -- connects. The picture wasn't as

:20:33. > :20:37.crisp and colourful as the Insta360. The camera is lightweight and the in

:20:38. > :20:40.battery didn't last long. But the three microphones offered good

:20:41. > :20:45.surround sound, something the will appreciate more if you what your

:20:46. > :20:49.movies through a VR headset. -- something you will. As the light

:20:50. > :20:53.faded, we decided to prepare the series kit that we would be using to

:20:54. > :21:00.fill out for high up on the mountain early the next morning. I brought

:21:01. > :21:07.the 360 Fly, which looks like a golf ball with an eye. That the camera

:21:08. > :21:12.with a 240 degrees superwide lens. That means there is no stitching

:21:13. > :21:19.together of shots and that should mean a smooth and clean picture. She

:21:20. > :21:24.used the Kodak double action camera. The two cameras need to be

:21:25. > :21:28.synchronised, so they are started by a remote-controlled watch so the

:21:29. > :21:31.record the same time. The image from the two cameras need to be stitched

:21:32. > :21:39.together later with Kodak software, if the stitching works well we

:21:40. > :21:46.should get winning results. We've been told Rwanda was stunning so we

:21:47. > :21:54.decided to trek 3000 metres up to take a look. A fellow adventurer at

:21:55. > :21:58.kindly agreed to be our cameraman, which means we strapped the golf

:21:59. > :22:02.ball to his head and it soon became apparent what the limitation of his

:22:03. > :22:10.single lens camera was. A great, lucky black pit at the bottom of the

:22:11. > :22:15.picture. Ones in the jungle it looked awful. To be fair, it can be

:22:16. > :22:20.cropped out later, leaving a better view that actually 360 horizontally

:22:21. > :22:24.but you can't look down. The superwide angle made everything

:22:25. > :22:30.seemed far away. Anything close up looked great, but the sound quality

:22:31. > :22:39.was ruined. As we trudged through the undergrowth, we decided it was

:22:40. > :22:45.time to swap over to the Kodak. It was then the adventure really took

:22:46. > :22:48.off. The air got thinner and this camera looked like it would capture

:22:49. > :23:01.anything we came across. Or anything that came across us.

:23:02. > :23:06.By having two super high-definition cameras we weren't just able to

:23:07. > :23:12.capture this incredible creatures wherever they went, but we have the

:23:13. > :23:16.resolution to zoom in as well. On the downside, the two cameras didn't

:23:17. > :23:21.automatically stitch well together. After fiddling with it using Kodak's

:23:22. > :23:26.on software, we decided on shot was running behind the other. After a

:23:27. > :23:33.calculated week got this much better results. -- tweak. The picture

:23:34. > :23:39.quality was the best of the bunch. The 360 cameras can allow you to

:23:40. > :23:43.capture everything in one go, but finer details still elude even the

:23:44. > :23:47.best of them, meaning it will still be a while before you feel like

:23:48. > :23:52.you're right there. That was Dan Simmons, clearly

:23:53. > :23:58.angling to be the 360 David Attenborough. That's it from the

:23:59. > :24:03.design Museum in London. Next week, it is the Click Christmas party, so

:24:04. > :24:07.be prepared for well, anything! Plus a look back at our best bits of

:24:08. > :24:33.2016. In the meantime, we live on Twitter. Thanks for watching!

:24:34. > :24:37.Friday was another grey day for many parts of the country. Any breaks in

:24:38. > :24:38.the cloud, mist and