:00:00. > :00:20.This week, wall climbing graffiti bots. All aboard the holodeck. And
:00:21. > :00:43.just watch out for the rodents. Oh, there is a mouth! -- mouse!
:00:44. > :00:55.Data is all around us. We generate around 2.5 billion GB of it every
:00:56. > :01:01.day. Think of it as, well, there is no other word for it, really.
:01:02. > :01:06.Enormous. And we are finding lots of new ways of gathering even more of
:01:07. > :01:11.it. Machines are now able to look at videos and interpret what is in the
:01:12. > :01:15.image, and with the number of CCTV cameras around the town, imagine how
:01:16. > :01:19.much more data we can collect. But the real intelligence is not in
:01:20. > :01:23.capturing the data. It is in analysing it. And this is where
:01:24. > :01:28.artificial intelligence might make a real difference, making connections
:01:29. > :01:31.that we humans never would. Big data has accelerated our understanding of
:01:32. > :01:37.medical science in unimaginable ways. It is now influencing how
:01:38. > :01:42.hospitals treat patients, police forces manage crime, and city
:01:43. > :01:47.officials run our towns. And it is inevitable, in the next 50 years,
:01:48. > :01:54.that AI will play an even bigger role in our society, and influence
:01:55. > :02:00.how we go about living. I recently met DJ Patil, President Barack
:02:01. > :02:04.Obama's chief data scientist, who was in charge of shaping how big
:02:05. > :02:08.data is used by the government to make big policy decisions, while
:02:09. > :02:12.ensuring the AIA created by the tech companies treat everyone fairly and
:02:13. > :02:16.make good decisions. That is where we have to start focusing more about
:02:17. > :02:19.energy, is asking the question of how do we actually make sure that
:02:20. > :02:23.these algorithms are going to work the way we want? People talk about
:02:24. > :02:28.self driving cars. It is a self driving car going to see someone
:02:29. > :02:33.with my skin tone, or someone with a darker skin tone? A person with a
:02:34. > :02:38.wheelchair? Is that a person in the dataset? How do start saying... You
:02:39. > :02:41.are suggesting whether a self driving car would recognise you as
:02:42. > :02:45.something it should avoid. Yes, Boyd, because we have different skin
:02:46. > :02:49.colour. Are people with your skin colour the only ones in the data
:02:50. > :02:54.set, am I ignored? Is that an accident? But what about somebody
:02:55. > :02:59.with a handicap? What about a kid on a tricycle? It is not sufficient to
:03:00. > :03:03.say oops about the algorithm. We have to figure out a more robust
:03:04. > :03:08.process as these things are becoming more integrated into our society.
:03:09. > :03:11.And if we have learned anything from this week's Facebook story, it is
:03:12. > :03:15.that tech companies are not the most transparent lunch. Facebook has been
:03:16. > :03:20.around for more than a decade, and only now, by chance, have we got a
:03:21. > :03:28.glimpse of how its moderators decide what we see on its platform. So how
:03:29. > :03:33.do we make sure the AI built by the same tech companies are using our
:03:34. > :03:37.data responsibly? So the first, it comes down to how are you trained?
:03:38. > :03:41.In our training these days, we often have found that technologists are no
:03:42. > :03:45.longer trained in humanities. One of the most critical components of
:03:46. > :03:48.humanities is the notion of ethics, so what we have called for is that
:03:49. > :03:52.every data scientist, every economist, anybody who works with
:03:53. > :03:55.data, must have ethics integrated throughout their entire curriculum.
:03:56. > :04:00.So you can start to have the conversation and dialogue about what
:04:01. > :04:03.are the ethical implications of the choices you make second part of this
:04:04. > :04:07.is how about security of the data? How do you make sure that you are
:04:08. > :04:10.actually building the algorithms with security, the datasets with the
:04:11. > :04:15.purity so people can't just break in. That has to no longer be elected
:04:16. > :04:18.or something outside. That has to be part of the core training. Once you
:04:19. > :04:22.have this component of our training, I think you are going to have a new
:04:23. > :04:26.set of people who have the vocabulary to talk about it. But
:04:27. > :04:29.that doesn't take into account the speed at which it is happening on
:04:30. > :04:32.taking place today. So what we do then? Number one, transparency.
:04:33. > :04:35.President Obama signed an executive order that said by default all data
:04:36. > :04:39.the Federal Government on the US Federal Government, publishers, must
:04:40. > :04:43.be open and machine-readable. And what that allows people to do is be
:04:44. > :04:49.able to access the data, preparing, use it, and innovate with it --
:04:50. > :04:53.compare it. And that is the problem, how do we strike the balance? We
:04:54. > :04:56.need to know that a AI system is not biased as loan from a dataset which
:04:57. > :04:59.includes all of us, and its decisions are fair, but we also
:05:00. > :05:03.don't want to stifle its process will make progress, because when it
:05:04. > :05:07.is used in the right way it really can change things for the better.
:05:08. > :05:11.What we have found in one of the problems around a local jail system
:05:12. > :05:15.is that there is a huge number of people who are just cycling in and
:05:16. > :05:17.out of the system. I mean, the numbers are extraordinary. More than
:05:18. > :05:22.11 million people through 3100 jails, they stay there on average 23
:05:23. > :05:26.days. 95% never go to long-term prison. It turns out there are a lot
:05:27. > :05:30.of mental health issues, a lot of drug addiction. So what happens to
:05:31. > :05:35.those people? Where is the data going? It stays in silos. The
:05:36. > :05:38.healthcare system has a silo, criminal justice. So what happens if
:05:39. > :05:46.you just talk and share that information? Is said do you see
:05:47. > :05:52.Sally and data set and say we saw Sally all the time? Well, why are we
:05:53. > :05:58.sending her to jail, let's send her to the right intervention. So doing
:05:59. > :06:02.that, how much can you save? What is the real impact? It costs 1.5
:06:03. > :06:06.million to train of Ron on the right intervention and share the data, and
:06:07. > :06:11.everything. The first year alone they saved more than $10 million,
:06:12. > :06:15.but more importantly they were able to close a full jail. And later on
:06:16. > :06:17.they close second jail because they are giving people the right care.
:06:18. > :06:23.It was the week that Volvo announced it's working on an AI rubbish truck
:06:24. > :06:26.that will follow collectors from house to house.
:06:27. > :06:29.IKEA said they will release smart light bulbs that can be controlled
:06:30. > :06:31.by voice and sync up with home devices like Alexa
:06:32. > :06:37.And Google fancied another "go" at Go success.
:06:38. > :06:40.The AI system AlphaGo took on the world's number
:06:41. > :06:44.one Go player Ke Jie, and won the series.
:06:45. > :06:47.AlphaGo learned to play by studying old matches and playing thousands
:06:48. > :06:53.The hope now is it will be used in medicine and science in the future.
:06:54. > :06:57.More bad news for Uber this week, as it admitted it underpaid drivers
:06:58. > :07:00.in New York for more than two and a half years.
:07:01. > :07:03.Tens of thousands of drivers will now be paid about $900 each,
:07:04. > :07:08.which will mean Uber paying out tens of millions of dollars.
:07:09. > :07:12.And only one month into the release of Samsung's new Galaxy S8
:07:13. > :07:16.smartphone virus scanner, and it's already been hacked.
:07:17. > :07:19.German hackers fooled the scanner with only a paper printer
:07:20. > :07:25.and a contact lens to make the fake eye.
:07:26. > :07:27.And is RoboCop from the '80s becoming a reality?
:07:28. > :07:31.Well, not quite, but Dubai police want these robots to make up 25%
:07:32. > :07:39.They launched the unit on Wednesday, which can forward video
:07:40. > :07:41.feeds to the police, settle fines, has facial
:07:42. > :07:43.recognition, and can speak nine languages.
:07:44. > :07:57.Graffiti art has been one of the hottest art movements over
:07:58. > :08:02.Like many graffiti artists, Graeme - or Xenz, the name he goes by -
:08:03. > :08:06.In this case, the streets of Bristol.
:08:07. > :08:10.And he has since grown into the artist that we see on the roof
:08:11. > :08:14.Today, he's taking a break to do this for us.
:08:15. > :08:18.But he's more known these days for these amazing natural scenes
:08:19. > :08:21.which are exhibited and sold all over the world,
:08:22. > :08:23.and which incorporate all of the graffiti techniques that
:08:24. > :08:29.Yeah, over time you really understand what the can
:08:30. > :08:34.You know, you come to rely on these tools, like the nozzle
:08:35. > :08:39.Like the way that I use the edge there to keep one edge
:08:40. > :08:41.sharp and one edge faded, then this, you know,
:08:42. > :08:49.So there's a lot of disciplines that go through painting that
:08:50. > :08:54.No, we don't have that kind of patience.
:08:55. > :08:57.So could we pull off something similar to this by combining
:08:58. > :08:59.technology with someone who has no creative talent whatsoever?
:09:00. > :09:02.To find out, we sent Nick Kwek to Estonia...
:09:03. > :09:03.Tartu, Estonia's second-largest city.
:09:04. > :09:05.Like most cities, graffiti and street art provoke
:09:06. > :09:24.It's also home to one of the biggest spray-painted pieces
:09:25. > :09:39.But Albert's been painted dot by dot, and I've been promised I too
:09:40. > :09:41.can achieve artistic genius with the right tools.
:09:42. > :09:44.Believe it or not, these pictures have all been
:09:45. > :09:49.They've been pieced together splodge by splodge
:09:50. > :10:00.My daughter wanted a unicorn on her wall, but I couldn't draw.
:10:01. > :10:03.So that pushed me towards creating this device.
:10:04. > :10:06.To make these magical masterpieces you need the right kit -
:10:07. > :10:08.a smartphone with the appropriate app installed, an external battery
:10:09. > :10:12.pack to keep it fully juiced, a tripod to hold it steady,
:10:13. > :10:29.some paint, and of course the SprayPrinter.
:10:30. > :10:32.First you select an image and align it against
:10:33. > :10:36.So the image is projected like a giant virtual sticker.
:10:37. > :10:40.The phone's camera exposes for the LED on the device,
:10:41. > :10:43.and when it illuminates it sends the can's location to the app.
:10:44. > :10:46.The phone then tells the printer its coordinates
:10:47. > :10:50.and the printer decides when to spray and when not to.
:10:51. > :10:53.Once you get the knack of it, it's actually surprisingly simple to use.
:10:54. > :10:57.You just have to make sure you go from left to right, or right to
:10:58. > :10:59.left, very smoothly, in a straight line.
:11:00. > :11:02.For all its geeky brilliance, it's a real labour of love.
:11:03. > :11:04.Even the most simple of designs takes several
:11:05. > :11:06.Depending on how complex the picture,
:11:07. > :11:08.and the size, the amount of
:11:09. > :11:10.layers, the different colours you want to paint with,
:11:11. > :11:13.you know, that determines how long doing one of
:11:14. > :11:26.You need to move your hand relatively
:11:27. > :11:28.steady, so if you start moving your hand very
:11:29. > :11:31.Not sure I could really stand your for
:11:32. > :11:38.With the next model, you should be able to
:11:39. > :11:41.move your hand relatively freely as you would with
:11:42. > :11:56.rest, the team have already started developing robotic
:11:57. > :11:59.versions to do the spraying for them, meaning larger more complex
:12:00. > :12:02.I developed this extra accessory for the SprayPrinter to
:12:03. > :12:04.atomise the process, because for high scale
:12:05. > :12:07.images the hand-held method takes too much time and
:12:08. > :12:18.too, hopefully speeding things up a bit.
:12:19. > :12:20.But does the printer help artistic expression, or
:12:21. > :12:27.gives like guidelines of how to paint.
:12:28. > :12:29.It's like sort of a colouring book, but
:12:30. > :12:31.you can go over the lines, but the paint
:12:32. > :12:33.will still only land in the
:12:34. > :12:46.I think for people like myself, we call them
:12:47. > :12:48.LAUGHTER And I think this device gives them
:12:49. > :12:52.It started off only a few small dots.
:12:53. > :12:55.You actually have to stand back a few feet to get the
:12:56. > :12:58.full view, to get the right perspective on it.
:12:59. > :13:12.So what would you like to see spray-painted next?
:13:13. > :13:15.Well, the guys have been holding a competition
:13:16. > :13:18.and this winning submission, just announced, will soon be painted on
:13:19. > :13:19.a local giant abandoned power station
:13:20. > :13:22.chimney for all to see, but painting on this
:13:23. > :13:23.curved structure has posed new
:13:24. > :13:25.problems, which Mihkel is determined to solve.
:13:26. > :13:29.I thought it would be a good idea to use a vacuum
:13:30. > :13:32.Rover, so this is just a four wheeled platform that drives across
:13:33. > :13:42.It attaches to the wall using vacuum.
:13:43. > :13:45.Yeah, and in true Blue Peter fashion, here's one I made
:13:46. > :13:59.Well, that was Nick Kwek with the SprayPrinter.
:14:00. > :14:09.It helps us to get these large images
:14:10. > :14:12.up easier but no, I think I'm quite comfortable
:14:13. > :14:17.It definitely has its advantages, for
:14:18. > :14:20.Well, in the meantime, this is beautiful.
:14:21. > :14:22.Thanks so much for doing this for us.
:14:23. > :14:29.We're going to stay on and art tip now.
:14:30. > :14:37.Here at Photo London art takes many forms.
:14:38. > :14:39.But the thing I've seen that I've grappled with the
:14:40. > :14:43.most is the idea of a virtual reality gallery.
:14:44. > :14:48.Is this really a way to fully experience art?
:14:49. > :14:57.So what's going on in here, and in here?
:14:58. > :15:00.Well, in the 1800s when people saw photography for the first time they
:15:01. > :15:03.were absolutely wowed by it, but of course now
:15:04. > :15:09.So what's happening is some of those initial images are being
:15:10. > :15:15.brought back to life in virtual reality.
:15:16. > :15:21.original original photographic images were shown has been recreated
:15:22. > :15:27.Well, initially I wasn't sure that looking at these images in
:15:28. > :15:30.virtual reality seemed like something that actually makes
:15:31. > :15:33.sense, but apparently you can pick up the
:15:34. > :15:37.images by holding your hand over it like that,
:15:38. > :15:40.and then you can hold the image in your hand...
:15:41. > :15:47.You can really see the texture of it as well.
:15:48. > :15:49.This genuinely feels like I'm standing in
:15:50. > :15:53.In fact, it actually feels quite hazardous
:15:54. > :15:57.because you can see smoke coming off it and that is proper serious heat.
:15:58. > :16:06.But whilst the juxtaposition between the origins of photography
:16:07. > :16:09.and a new visual medium are deliberate, making sure it
:16:10. > :16:13.provides a meaningful experience for those with a yearning for art
:16:14. > :16:21.Nothing fills me with a greater melancholy than going
:16:22. > :16:24.into an exhibition and seeing somebody with a virtual reality
:16:25. > :16:28.headset on, and having to queue and wait for your turn on it -
:16:29. > :16:33.So what I've tried to do in this installation is to make that part
:16:34. > :16:36.of the actual experience, so when you're not in the room
:16:37. > :16:38.you can look at people with their headsets
:16:39. > :16:42.Watching the goings-on of people wandering around is quite strange
:16:43. > :16:45.and surreal to look at, so hopefully it's still interesting
:16:46. > :16:49.as an artwork even when you're not in the headset.
:16:50. > :16:52.So I can hear some sound coming from over here.
:16:53. > :16:54.That's because of the binaural sound that's built in,
:16:55. > :16:59.and there seems to be something happening outside...
:17:00. > :17:02.I believe this is the Chartists' revolt.
:17:03. > :17:06.This is a lot of people objecting to photography.
:17:07. > :17:09.This wasn't the only VR at the show, though.
:17:10. > :17:11.One family of art collectors wanted to virtually take
:17:12. > :17:24.You can have your art museum in your pocket.
:17:25. > :17:26.I can have 200 metre museum just in my laptop.
:17:27. > :17:29.That could be sharing a collection internationally, a trip to a
:17:30. > :17:32.virtual art gallery for those who are housebound, or
:17:33. > :17:35.introducing a new audience to art who might be more
:17:36. > :17:38.The real-life version of this statue is
:17:39. > :17:42.I will head towards it and have a closer look.
:17:43. > :17:44.I can actually see the size of it by those
:17:45. > :17:48.And in fact the size of that piece of art behind
:17:49. > :17:51.it, the scale of all of this, is absolutely massive.
:17:52. > :17:54.It would require such a large building to actually
:17:55. > :17:59.Amidst the physical art were the latest
:18:00. > :18:03.imaging, entire film is superimposed on
:18:04. > :18:07.single images, and this Paris park scene.
:18:08. > :18:10.So behind this photograph we are looking at here is actually a
:18:11. > :18:14.massive plate of LED lights, all spread out with an inch between
:18:15. > :18:21.them, so each time you can see a person crossing the screen it's
:18:22. > :18:24.actually a combination of these lights being dimmed in that pattern,
:18:25. > :18:29.and what the human eye fills an in between to make it
:18:30. > :18:34.One thing that seemed clear by the end of the day,
:18:35. > :18:38.though, was that VR can feel a natural part of an art show, and
:18:39. > :18:42.that I'm never going to be an art expert.
:18:43. > :18:56.One of the brilliant things about working
:18:57. > :19:01.ambitions at one point or another, which is why
:19:02. > :19:06.this week Mark Cieslak became the captain of a starship!
:19:07. > :19:09.He took some of the rest of the Click family
:19:10. > :19:14.with him, to boldly go where no Mark has gone before.
:19:15. > :19:19.These are the virtual voyages of the BBC Click
:19:20. > :19:25.Our mission: To wear VR headsets and discover strange, new
:19:26. > :19:50.technology, and boldly go where no TV reporter has gone before.
:19:51. > :19:54.Virtual reality game Star Trek Bridge Crew
:19:55. > :19:55.brings together up to four players, each
:19:56. > :19:57.taking a different role on the
:19:58. > :20:03.The beauty of going where no one has gone before
:20:04. > :20:07.is that starship travel involves an awful lot of sitting down.
:20:08. > :20:09.Sitting down is great for virtual reality
:20:10. > :20:12.because the headsets have got these cables.
:20:13. > :20:16.If you're moving around it easy to get caught up with them.
:20:17. > :20:24.And where better to be sitting in the
:20:25. > :20:26.Captain's chair of a Federation starship?
:20:27. > :20:29.Headsets on, it's time for the Click team to become a starship
:20:30. > :20:54.The early missions are all about orientating us with the bridge
:20:55. > :20:58.As helmsman, you are the ship's navigator.
:20:59. > :21:01.The headset shows us what the bridge looks like, but the
:21:02. > :21:03.PlayStation motion controllers allow us to interact with the various
:21:04. > :21:07.controls we have to master in order to fly the ship.
:21:08. > :21:29.We don't have time for sight seeing, though, as we receive
:21:30. > :21:31.a distress signal from a stricken vessel.
:21:32. > :21:34.My vessel has lost all power and our life-support systems are
:21:35. > :21:44.Can you transport the survivors to here?
:21:45. > :21:54.CHUCKLES That wasn't in the training.
:21:55. > :21:56.LAUGHTER We're homing in at an alarming rate,
:21:57. > :21:58.There are no options within transporter.
:21:59. > :22:07.It's at this moment that the action takes
:22:08. > :22:11.a turn which will appeal to Star Trek superfans.
:22:12. > :22:16.OK, guys, this is the Kobayashi Maru scenario.
:22:17. > :22:19.This is an impossible to win situation.
:22:20. > :22:41.Bring us about so we can actually see that
:22:42. > :22:56.Line up the phasers, and torpedoes away.
:22:57. > :23:06.Yeah, everybody, we just violated a peace treaty.
:23:07. > :23:11.It's pretty warm work being in virtual reality.
:23:12. > :23:14.It feels like it's social VR at its best, really.
:23:15. > :23:18.If you don't have it you're not going to complete the mission.
:23:19. > :23:22.I thought we actually had our lives on the
:23:23. > :23:25.That ably demonstrates the power of teamwork.
:23:26. > :23:28.It's really, really important that everybody plays their role on the
:23:29. > :23:31.bridge, because if you don't then chaos ensues.
:23:32. > :23:34.We had a couple of sticky moments there, but I think we
:23:35. > :23:38.managed to pull it back and keep it together as a crew.
:23:39. > :23:41.And the result was a successful mission.
:23:42. > :23:50.Or like us on Facebook, too, where you can see
:23:51. > :23:55.Now, while you're watching this we are doing a live show at the Hay
:23:56. > :24:02.And next week on the programme you can see a
:24:03. > :24:07.little bit of what we're getting up to.
:24:08. > :24:10.And if you're coming, I hope you enjoy the
:24:11. > :24:36.With a bank holiday weekend now upon us,
:24:37. > :24:39.we are set to see a change in the hot, dry weather,
:24:40. > :24:42.that has been with us for the past few days.
:24:43. > :24:44.Here was the scene on Friday in Moray.
:24:45. > :24:47.One of our Weather Watchers captured this.