Do Androids Dream of Chocolate Cake?

Download Subtitles

Transcript

:00:00. > :00:00.Now on BBC News, it's time for Click.

:00:07. > :00:19.This week, destressing with a future ball, knocking up Google, and

:00:20. > :00:49.shouting at Amazon. Tell me, who is the murderer! THEME SONG.

:00:50. > :00:59.Welcome home. How was your day? Awful. That is a shame. I will run

:01:00. > :01:14.you a bath and play relaxing music. MUSIC PLAYS. Rory, do I have any

:01:15. > :01:17.messages. 17,000 Tweets, and fake news updates. Anything urgent? Your

:01:18. > :01:29.boss wants dinner tonight. Is there anything to eat? A scallop which

:01:30. > :01:32.feeds four. The oven will come on now and his favourite wine will come

:01:33. > :01:36.at seven. And get some chocolate double fudge cake. I have ordered

:01:37. > :01:42.it. Shall I apply for a gym membership for you? Mute. One day we

:01:43. > :01:46.really will have artificial intelligence in our personal

:01:47. > :01:54.assistance which we can talk to and who know us better than ourselves.

:01:55. > :02:00.Thank you, Rory. You are welcome. We are not there yet, but we are well

:02:01. > :02:15.on the way. What started in our phones with names like Siri,

:02:16. > :02:19.Cortana, and, uh, OK, Google, can now control cars and homes too. Echo

:02:20. > :02:24.led the way. And Google is coming out in the UK. It is good to have

:02:25. > :02:34.personal assistants, but the more we use them, the more trust we have to

:02:35. > :02:40.placed in them. OK, Google, is Obama planning a coup? According to

:02:41. > :02:45.secret. In his day job, Rory recently discovered that you cannot

:02:46. > :02:52.always believe what they say. Obama is going to have a communist who at

:02:53. > :02:56.the end of his term in 2018. That fake news story was the top search

:02:57. > :03:00.result for that question. Dan Simmons has been looking at some of

:03:01. > :03:05.the other unintended consequences of living with these devices.

:03:06. > :03:14.As we transition from controlling things from screens to using our

:03:15. > :03:20.voice, those providing services could start getting tricky. I am in

:03:21. > :03:25.the BBC's Blue Room, a space where the broadcaster posts out new

:03:26. > :03:29.technology. And with the voice assistance, it is not all going

:03:30. > :03:34.smoothly. Alexa, when is the next train to Manchester? Sorry, I did

:03:35. > :03:39.not understand the question I heard. If you have to find out when the

:03:40. > :03:47.next train to Manchester is, right now you have to say, open the

:03:48. > :03:51.National Rail App and go through other steps. That is not natural.

:03:52. > :03:55.You have to go through many steps to find out content from somebody else

:03:56. > :04:00.to be the default service provider is very simple. Blame a song from a

:04:01. > :04:05.certain provider, tell me the news from a certain provider. -- play me

:04:06. > :04:09.a. That is a great thing for that, but everything else is critical. A

:04:10. > :04:14.lot of work needs to be done to level the playing field. And that

:04:15. > :04:19.disadvantage applies to search results as well. Up until now,

:04:20. > :04:24.website and to be on the first page of results. With voice assistance,

:04:25. > :04:31.just one answer comes back. OK, Google, how far is the moon? That is

:04:32. > :04:36.fine if it is a right or wrong definitive answer. That is what

:04:37. > :04:40.companies do most. It is this far from Earth. More controversial if

:04:41. > :04:45.you are looking for a product or service. For anyone else, how did

:04:46. > :04:48.you get to that position? Only one person can have the first spot.

:04:49. > :04:55.Everyone else will have to figure out what did they do, how do they

:04:56. > :05:01.work with Amazons and Googles to figure out how to get their results

:05:02. > :05:05.first. This is not the end of the world, it is just the end of

:05:06. > :05:08.competition as we know it. Oxford University is home to one of the

:05:09. > :05:14.world's most influential thinkers when it comes to competition to be

:05:15. > :05:18.if we use our assistants to buy stuff, this man believes there will

:05:19. > :05:23.be consequences, and they will not be unintended ones. That shift from

:05:24. > :05:27.that environment to the digital helper, what is it that you have?

:05:28. > :05:33.You have a helper that is voice-activated, you are one step

:05:34. > :05:36.further from the ability to look for outside options. Your ability to

:05:37. > :05:42.check whether the price you received is truly the best price. You tell

:05:43. > :05:45.your helper, order me one, two, three, and you just assumed that the

:05:46. > :05:49.helper will serve your needs. The likelihood is that in a two side

:05:50. > :05:55.market, the helper is actually serving the platform. Today, your

:05:56. > :06:02.assumption, our default assumption, is the price you receive is the

:06:03. > :06:07.competitive rice. -- price. And you are suggesting that it will not be?

:06:08. > :06:11.I am telling you that it is not. A walk down Oxford's Cornmarket street

:06:12. > :06:17.reveal something the professor believes will not be around much

:06:18. > :06:22.longer on line. How much is this, for example? How much is this today?

:06:23. > :06:28.?5. This gentle man over here, hello, how much would you sell this

:06:29. > :06:33.to him, and how much to me? You just met us in the street. The same price

:06:34. > :06:40.for everyone. Of course. Absolutely. Do you think I would pay more? No.

:06:41. > :06:48.To be honest, lately, tourists by more. Really? You think you made for

:06:49. > :06:58.-- may pay more than me? It is one price for everyone, but digital

:06:59. > :07:01.assistants will get to know us so well, prices will be tailored to us

:07:02. > :07:07.as well, effectively becoming a gatekeeper to the best deals. I went

:07:08. > :07:11.to see one of those gatekeepers, Google, and asked them if sellers

:07:12. > :07:18.goodbye their way to the top result and get recommended by their digital

:07:19. > :07:23.assistant. -- could purchase. The consumer is the main focus for us.

:07:24. > :07:27.Doing something like that will not help them find what they want. We

:07:28. > :07:32.want to make sure we are focused on what they want. Amazon told us there

:07:33. > :07:36.is lots of potential and room for many participants to be our job is

:07:37. > :07:42.to innovate on behalf of the customer and then let customers

:07:43. > :07:46.decide. By perhaps what these home assistants are most useful for is

:07:47. > :07:52.what they are becoming most known for, and that is controlling other

:07:53. > :08:17.things around the house. Alexa, turn on the bar lights. OK. Alexa, bar

:08:18. > :08:23.lights off. OK. PHONE RINGS. Hi. Dan, are you there? Look, I know we

:08:24. > :08:29.have not seen each other, and you think I am crazy, but I was just

:08:30. > :08:39.passing by, and... Oh, wait, have you still got that stupid voice

:08:40. > :08:51.control thing, what was it, Alexa? Turn on the bar lights. OK. Alexa,

:08:52. > :08:57.turn on the microwave. Have I got your attention now? Alexa, unlock

:08:58. > :09:03.the front door. OK. It is only me. We set that up. But the lights were

:09:04. > :09:08.real, even though the other and in the front door was faked a little

:09:09. > :09:12.bit by us to just show you what the potential is of this technology if

:09:13. > :09:22.it cannot recognise your voice. In actual fact, Amazon tell us the

:09:23. > :09:25.unlock feature for doors is not available on the Echo, and that may

:09:26. > :09:32.be the biggest admission there is that there is a lot to be done with

:09:33. > :09:42.security on these devices. Welcome to the Week in Tech. . The week that

:09:43. > :09:50.Apple said its latest apps laptop was a bad design. And Blizzard, the

:09:51. > :09:58.maker of Overwatch, successfully sued a cheater for copyright

:09:59. > :10:04.infringement. But Graphine stole the show. They have created a means of

:10:05. > :10:09.removing salt from sea water, eventually bringing potential to

:10:10. > :10:18.provide cheap water. Amazon has been selling a billion dollars' worth of

:10:19. > :10:22.shares a year for a space project. They had to send travellers into

:10:23. > :10:25.space in the next two years. In Massachusetts, a robotic arm picks

:10:26. > :10:29.up random objects and puts them on a conveyor belt all day and shares its

:10:30. > :10:34.information with other robots so they can do it too. It could be a

:10:35. > :10:43.warehouse of the future. And a man flew with a robot arm developed in

:10:44. > :10:48.his garage. It has six jet engines. It can fly hundreds of miles per

:10:49. > :11:03.hour. Although Richard is exercising restraint, he said, for now. Amazon

:11:04. > :11:07.Echo, Google Home, maybe one day even Rory. It looks like the rise of

:11:08. > :11:12.the digital assistant is upon us. All of these are trying to be wide

:11:13. > :11:16.ranging all-purpose artificial intelligence systems. That is the

:11:17. > :11:22.technical term, Horizontal AI. That is hard. To cover a lot of subjects

:11:23. > :11:29.and too many tasks, these things have to understand a lot of things.

:11:30. > :11:34.Rory, tell Stephen I will call him in the office after my train

:11:35. > :11:38.journeys. First it works out what I said. Then it pulls out the

:11:39. > :11:44.important words. Then comes the reasoning and context it needs, and

:11:45. > :12:03.it does so by scanning my calendar and my train timetable and guess is

:12:04. > :12:07.that he is the Stephen in the office, not the friend, and then the

:12:08. > :12:11.right action needs to be performed, schedule the call and let him know.

:12:12. > :12:14.But ROVs best placed to do that? -- are these. Or should we speak to

:12:15. > :12:21.other specialist Vertical AIs that it do one job really, really well. X

:12:22. > :12:29.dot AI is a company that thinks that. We scheduled a call for his

:12:30. > :12:32.secretary, Andrew. It was only after five messages we realised that

:12:33. > :12:38.Andrew is actually artificial intelligence, a specialist AI that

:12:39. > :12:43.only scheduled meeting. I have to say it interpreted Aaron messages

:12:44. > :12:48.and request pretty well. Shore enough, on line is the man who

:12:49. > :12:59.invented it, Dennis. Congratulations, you fooled us with

:13:00. > :13:02.Andrew. Will we have any general assistants in the future or many

:13:03. > :13:26.vertical ones? We imagine this will play at in the short-term future.

:13:27. > :13:29.Think Siri, Alexa, so on, they will answer simple questions, but will

:13:30. > :13:32.also enable us to talk to vertical AIs which are specialised in doing

:13:33. > :13:51.one job really well. I think what will happen is you will have a whole

:13:52. > :13:54.plethora of Vertical AIs on your payroll, like Siri and Cortana, with

:13:55. > :13:58.13 agents on payrolls to do jobs you don't want to do. There is a whole

:13:59. > :14:01.application store for this with different jobs in different needs.

:14:02. > :14:04.It will be the same with intelligent agents. You will have different

:14:05. > :14:15.needs. Take you for your time. Sender Aaron love to Andrew. -- Send

:14:16. > :14:22.our love. In San Francisco, we have been living a very smart life.

:14:23. > :14:31.Set the temperature to 72 degrees. Setting to Brita to 72 degrees... To

:14:32. > :14:35.be gay I is a faithful servant arranger Rome. Its control comes

:14:36. > :14:39.from and read out, and the Nicolet proud of the way it understands when

:14:40. > :14:46.you are saying. Turn the lights on in the living room. Turning light on

:14:47. > :14:52.in the living room. Make it read. Setting colour to read in the living

:14:53. > :14:54.room. Typically, when you have different devices from different

:14:55. > :14:58.brands and manufacturers that have a bunch of apps for every device, and

:14:59. > :15:02.it is really inconvenient to jump around and switch from one up to

:15:03. > :15:12.another. The app integrates with other popular home devices. But like

:15:13. > :15:16.most smart home assistance, the commands are pretty basic. In the

:15:17. > :15:20.future, we will add scenes. When you wake up in the morning, you can say

:15:21. > :15:23.good morning, and then Peter Richel will be comfortable view and the

:15:24. > :15:28.lights will be on in certain areas. These home assistance are still a

:15:29. > :15:32.little buggy and frustrating at times. But they are getting there.

:15:33. > :15:38.And study to feel quite useful. The next challenge is to take those

:15:39. > :15:42.assistance outside of the home. Ford recently opened this research centre

:15:43. > :15:47.in the silicon valley. One of its main project is to integrate smart

:15:48. > :15:53.assistance into its cars. They have been working with others on to work

:15:54. > :16:01.in the Alexa assistant. You can find out important information about your

:16:02. > :16:05.car. Moto3, asked me ford mobile for my tyre pressure. Your vehicle's

:16:06. > :16:16.tyre pressure is not currently showing any warnings. -- Alexa, show

:16:17. > :16:20.me my Ford. If you had to wait in a freezing vehicle as it warms up, you

:16:21. > :16:29.might appreciate this. Ask me Ford to warm up my car. OK, so your pin.

:16:30. > :16:42.Five, six, seven, eight. Sending start command to the car. I mean, it

:16:43. > :16:46.is not quite the roar of the engine, but it is surely be causing you can

:16:47. > :16:49.do with this technology right now. Out on the road, the assistant steps

:16:50. > :16:58.in to make typical in car function of little more hands-free. --

:16:59. > :17:04.functions. Alexa, continue reading my audio book. OK, so I can pick up

:17:05. > :17:13.where I left off. The three, find the nearest coffee shop. Here are a

:17:14. > :17:17.few nearby options... The integration is fun, but far from

:17:18. > :17:21.perfect. Alexa still suffers from the same problem that many

:17:22. > :17:24.assistants have, rather than talking naturally to it, you find yourself

:17:25. > :17:27.having to think about what phrase will unlock the information you

:17:28. > :17:34.need. When you're trying to drive, that silicon could quite

:17:35. > :17:37.distracting. It has a leg which detection system, so does understand

:17:38. > :17:41.what you say, independent of how you say it. I wonder if people will be

:17:42. > :17:45.thinking more about Alexa is thinking, rather than what is ahead

:17:46. > :17:49.of them on the road. As a driver, we want to make driving safer, so you

:17:50. > :17:54.should always keep your hands on the steering wheel and focus on the

:17:55. > :17:57.road. And then to your entertainment and entertainment on the road, you

:17:58. > :18:08.could use your voice, which is really the safest way to interact

:18:09. > :18:17.with the car, in general. James looked at it different in the early

:18:18. > :18:23.80s. To get Dolby 5.1 audio. To get realistic lighting effects. Forget

:18:24. > :18:28.about 3-D graphics in most cases. Attitudes to games won't exactly

:18:29. > :18:33.enthusiastic, either. Sports games. Well, these are perhaps the most

:18:34. > :18:40.appalling list of computers to me. This is to cap on. Doing a decathlon

:18:41. > :18:47.on a Commodore, white wagging a joystick, is not a substitute for

:18:48. > :18:51.fresh air. Ouch. There was one genre of games, though, where the graphics

:18:52. > :18:55.and audio did not matter. Where did not matter that you would often be

:18:56. > :18:59.presented with a still, 8-bit image, or sometimes just a black screen

:19:00. > :19:04.with a flashing white coaster. That screen itself was the window into

:19:05. > :19:10.worlds of limitless imagination. Welcome to the experience of the

:19:11. > :19:14.text adventure. When competing power is limited, the text adventure that

:19:15. > :19:20.players head scratching puzzles and mysteries, all brought to life by

:19:21. > :19:25.typing instructions into the game. But the reason that I have taken us

:19:26. > :19:29.on this journey down 32 kilobits memory lane is because of the games

:19:30. > :19:34.that have been playing on this, the Amazon Echo. It is titled I see

:19:35. > :19:47.those old text adventures quite a bit. Leading you through the Abbey,

:19:48. > :19:55.the bass approach is one of the sisters. -- abbess. You might think

:19:56. > :19:59.playing a game in one of these is like trying to play a game on a

:20:00. > :20:03.microwave. Because it lacks a screen or any other way of interacting with

:20:04. > :20:11.it other than barking commands at it, but that is exactly how the game

:20:12. > :20:25.I am about to play works. Play risque. -- Runescape. This story is

:20:26. > :20:29.about an assassin Demon... The player must solve a murder in a

:20:30. > :20:35.fantasy realm. The gameplay is like an interaction to make interactive

:20:36. > :20:36.version of an audio book. You get on dialogue, then it went for a

:20:37. > :20:52.response. Surprisingly, it commands quite a

:20:53. > :20:56.bit of your attention, and it is quite a relaxing way to play a game,

:20:57. > :21:01.although that relaxing mood is shattered when you hear this. Sorry,

:21:02. > :21:12.that is not a valid command. Which you hear quite a lot. Would you like

:21:13. > :21:23.to talk to the abbess are now? Talk to the abbess. Sorry... Talk to the

:21:24. > :21:27.grocery... Sorry, that is not a valid command. -- talk to the

:21:28. > :21:31.abbess. It can shatter the elution and become increasingly frustration

:21:32. > :21:34.when it does not understand what you are saying. Which is busily bit of a

:21:35. > :21:38.problem for again you play by talking to it. Sorry, that is not a

:21:39. > :22:00.valid command. When it does work, though, Runescape

:22:01. > :22:03.on the Amazon Echo is fun and immersive. It highlights the

:22:04. > :22:07.potential these devices have beyond reading at the weather to you or

:22:08. > :22:17.reciting rubbish jokes. Runescape is available by the skill section of

:22:18. > :22:23.the Alexa up. -- app. Have a sick bay filled with headaches... When

:22:24. > :22:27.Star Trek introduced a device to scan a passionate and, with a

:22:28. > :22:31.diagnosis, it was in the realm of science fiction. But 50 years on,

:22:32. > :22:35.this dream is becoming a reality. It, we will discover the winner of

:22:36. > :22:41.$18 million prize playing a march to such a's medical device. The

:22:42. > :22:52.challenge? To design and build a device that will register 13

:22:53. > :22:55.conditions. It could and visits to the doctor. We figured out what the

:22:56. > :23:03.diagnostic process is, at least the way I am doing it, and build a

:23:04. > :23:09.system with that in mind. So it is not like a single device like the

:23:10. > :23:16.Star Trek series. It lets us you interact with it on a tablet in our

:23:17. > :23:19.prototype, and then plugged in components. So we have a little

:23:20. > :23:28.device like this. User will be guided, they are guided to listen to

:23:29. > :23:38.sounds. Des Connor device in your home is really a medical centre

:23:39. > :23:42.right in your house. -- this kind of device. If your child is ill, if you

:23:43. > :23:48.are not feeling well, it is your first stop. There is a revolution

:23:49. > :23:54.coming in healthcare. This is the type of device that is going to help

:23:55. > :24:02.give people the power to take care of themselves. And as soon as we

:24:03. > :24:05.know who has one, we will let you know on Twitter. Think you for all

:24:06. > :24:10.your interactions on Twitter, too, which this week included choosing

:24:11. > :24:17.the name of our artificial intelligent to intelligence. So good

:24:18. > :24:21.night, Rory. See you soon.