Return to Video

Blab Droids -- Self-Learning Social Cloud Robots: Alex Reben at TEDxBeaconStreet

  • 0:16 - 0:19
    So, we're becoming evermore
    integrated with technology.
  • 0:19 - 0:20
    It's everywhere.
  • 0:20 - 0:22
    But technology is really
    what we make of it.
  • 0:23 - 0:26
    And, we should really
    bring it to the next level.
  • 0:27 - 0:29
    And I think the way we can do this,
  • 0:29 - 0:32
    is ask the question,
    "how can technology learn from us?"
  • 0:32 - 0:34
    And the way it can learn from us,
  • 0:34 - 0:35
    is that we can teach it.
  • 0:35 - 0:36
    We can teach about humanity.
  • 0:36 - 0:39
    And by doing so,
    it can make us more human.
  • 0:39 - 0:42
    Lack of understanding
    knowledge and empathy,
  • 0:42 - 0:45
    is the root cause
    of many human problems.
  • 0:45 - 0:47
    And I believe we could use technology,
  • 0:47 - 0:49
    to increase our level of understanding.
  • 0:49 - 0:51
    We can get knowledge from places
  • 0:51 - 0:53
    like the Internet,
    but the Internet in itself
  • 0:53 - 0:55
    is not inquisitive.
  • 0:56 - 0:57
    These are the BlabDroids.
  • 0:57 - 1:01
    And they are robots created
    to learn about people.
  • 1:01 - 1:03
    And they do this by going around,
  • 1:03 - 1:05
    and asking people questions,
  • 1:05 - 1:08
    as most of us do when we want to learn.
  • 1:08 - 1:10
    And you can see they are very cute,
  • 1:10 - 1:13
    and made to be very friendly.
  • 1:13 - 1:15
    Because if you want to have
    a relationship with a robot,
  • 1:15 - 1:16
    it better be pretty nice .
  • 1:16 - 1:18
    (Laughter)
  • 1:18 - 1:20
    So, next you'll see a couple of examples
  • 1:20 - 1:21
    of the BlabDroids asking some people
  • 1:21 - 1:24
    somewhat difficult questions.
  • 1:25 - 1:27
    Robot: if you could give someone any gift,
  • 1:27 - 1:29
    what would it be?
  • 1:29 - 1:32
    Woman: Give my mother the gift
  • 1:32 - 1:35
    of not worrying about me
  • 1:36 - 1:37
    before she dies.
  • 1:38 - 1:40
    And she wants me to lose,
  • 1:40 - 1:41
    like a ton of weight,
  • 1:41 - 1:43
    and get really really healthy.
  • 1:44 - 1:46
    And she needs to see that,
  • 1:47 - 1:48
    before she dies,
  • 1:48 - 1:52
    For her to feel like I am going
    to be Ok when she is not here,
  • 1:52 - 1:54
    and I wish I could give her that.
  • 1:54 - 1:56
    And I am not positive I can.
  • 1:58 - 1:59
    You asked!
  • 2:01 - 2:04
    Robot: if you died tomorrow,
    what would you regret the most?
  • 2:05 - 2:08
    Woman: I switched out her shampoo
  • 2:08 - 2:11
    for a Nair hair remover
  • 2:11 - 2:12
    when she wasn't looking,
  • 2:12 - 2:16
    and clumps of her hair fell out.
  • 2:16 - 2:18
    I felt like she deserved it,
  • 2:18 - 2:20
    it's very much still justified .
  • 2:21 - 2:23
    And I would do it again to her
  • 2:23 - 2:24
    in a heartbeat.
  • 2:24 - 2:25
    I still don't like her.
  • 2:25 - 2:26
    So these are the types of things
  • 2:26 - 2:28
    that you may not tell a person
  • 2:28 - 2:31
    who is coming up to you with questions,
  • 2:31 - 2:36
    but the robots in fact got those answers.
  • 2:36 - 2:37
    So, the ideas are these robots will become
  • 2:37 - 2:42
    a tangible manifestation
    of cloud intelligence
  • 2:42 - 2:44
    that can both learn and teach.
  • 2:44 - 2:45
    And as I said before,
  • 2:45 - 2:48
    the Internet in itself
    doesn't necessarily learn,
  • 2:49 - 2:51
    but these robots can be made in a way
  • 2:51 - 2:55
    that we want to interact with them
    on a very high level.
  • 2:55 - 2:58
    Think of them as if you had
    a dog in your house
  • 2:58 - 2:59
    that asked you questions,
  • 2:59 - 3:02
    but you could also ask it questions.
  • 3:03 - 3:04
    So how do we make
    something like this happen?
  • 3:04 - 3:07
    Well the first step is making it,
  • 3:07 - 3:08
    and we partnered with a few people like
  • 3:08 - 3:10
    Dragon innovation to do manufacturing
  • 3:10 - 3:12
    and bolt to form a company around us.
  • 3:14 - 3:15
    The idea is that these Robots,
  • 3:15 - 3:18
    will connect to the cloud
    through your cell phone
  • 3:18 - 3:19
    using bluetooth.
  • 3:19 - 3:21
    So we're all carrying around
  • 3:21 - 3:22
    these devices in our pockets
  • 3:22 - 3:24
    that are basically very smart computers,
  • 3:24 - 3:26
    and they also have network connectivity.
  • 3:27 - 3:29
    So, we can do things with a robot like,
  • 3:29 - 3:30
    asking simple questions,
  • 3:30 - 3:31
    you know, will it rain?
  • 3:31 - 3:33
    The robot would say, yeah, it might rain,
  • 3:33 - 3:35
    you may want an umbrella.
  • 3:35 - 3:37
    But asking it more
    difficult questions like,
  • 3:37 - 3:38
    are we alone in the universe,
  • 3:38 - 3:40
    or things of that sort,
  • 3:40 - 3:41
    Need a bit more of a human touch
  • 3:41 - 3:46
    than things like Google or Wolfram Alpha,
  • 3:46 - 3:49
    or things of that sort,
    may have a hard time answering.
  • 3:49 - 3:50
    And one way to do this
  • 3:50 - 3:52
    is use cloud knowledge
    and cloud intelligence.
  • 3:52 - 3:55
    So say you have
    a question for BlabDroid,
  • 3:55 - 3:57
    and it isn't one of
    the simple things like,
  • 3:57 - 4:00
    do I have a meeting
    scheduled for tomorrow?
  • 4:00 - 4:03
    It might say, "Well, gee!
    I don't know I'll ask my friends."
  • 4:03 - 4:05
    So the robot then accesses the cloud,
  • 4:05 - 4:07
    and talks to all his other
    BlabDroid buddies
  • 4:08 - 4:10
    and asks them, "hey,
    does your human know
  • 4:10 - 4:12
    something about this person's question?"
  • 4:12 - 4:15
    They may say, "Yes, let me contact them."
  • 4:15 - 4:17
    But as you can see,
    some of these guys are a little busy
  • 4:17 - 4:19
    but a couple of them
    may have a good idea.
  • 4:19 - 4:23
    So these people can
    tell the robot the answer,
  • 4:23 - 4:25
    either by looking up or
    maybe they're an expert in the field.
  • 4:25 - 4:28
    And that knowledge
    can be returned to that robot.
  • 4:28 - 4:32
    And at the same time,
    that knowledge can be saved online
  • 4:32 - 4:35
    in this sort of emergent intelligence
  • 4:34 - 4:38
    where, OK, the robot couldn't find it
    through traditional means,
  • 4:38 - 4:40
    now it's stored that human knowledge.
  • 4:40 - 4:43
    So imagine like,
    a Robot asking grandma
  • 4:43 - 4:45
    what her favorite cookie recipe is.
  • 4:45 - 4:47
    That's something that might have been
    lost to time in the past.
  • 4:47 - 4:52
    But by using these social
    interfaces that can be stored.
  • 4:52 - 4:54
    And there are obviously things right now
  • 4:54 - 4:55
    where you can talk to your cell phone,
  • 4:55 - 4:57
    and talk to other things where
  • 4:57 - 4:59
    you can ask it questions
    and they'll give you answers,
  • 4:59 - 5:01
    but the primitive parts for our brains
  • 5:01 - 5:03
    don't think cell phones are alive,
  • 5:03 - 5:04
    even if they had voices because
  • 5:04 - 5:07
    they're not persistent beings,
  • 5:07 - 5:10
    they're not objects that always stay
  • 5:10 - 5:15
    alive, in a way, where, a robot,
    a robot in itself is a physical form,
  • 5:15 - 5:18
    so even though it is connected
    to this cloud intelligence,
  • 5:18 - 5:22
    it's something that our brains
    perceive as alive,
  • 5:22 - 5:24
    and the way you want to do this,
  • 5:24 - 5:27
    is that we want them
    to be very fun and playful,
  • 5:27 - 5:28
    so you want to talk to them,
  • 5:28 - 5:29
    you want to converse with them,
  • 5:29 - 5:31
    not only asking them things,
  • 5:31 - 5:35
    but also having them
    with you as companions.
  • 5:35 - 5:39
    So in the future,
    we think that these can create a...
  • 5:41 - 5:43
    until now undocumented part of humanity.
  • 5:43 - 5:44
    So a lot of this oral tradition,
  • 5:44 - 5:45
    a lot of storytelling,
  • 5:45 - 5:47
    which has been, interacting
    into humanity,
  • 5:47 - 5:49
    has been kind of lost in the Internet age
  • 5:49 - 5:52
    for these can start getting ideas.
  • 5:52 - 5:54
    Reduce costs of cell phones.
  • 5:54 - 5:57
    will allow these to permeate more places,
  • 5:57 - 5:59
    And a new breed of social knowledge
  • 5:59 - 6:02
    through sharing and interacting
    allows to connect to people,
  • 6:02 - 6:07
    that maybe culturally disperse.
  • 6:07 - 6:08
    There are a lot of problems
    to solve obviously,
  • 6:08 - 6:10
    technical issues such as
  • 6:10 - 6:12
    voice recognition and data storage,
  • 6:12 - 6:14
    and that AI in the backend,
  • 6:14 - 6:16
    issues such as privacy and security,
  • 6:16 - 6:19
    Because obviously
    you're telling the robot things,
  • 6:19 - 6:23
    and other problems such as
    will personal robots be adopted?
  • 6:23 - 6:26
    That's more of a social question.
  • 6:26 - 6:29
    So to trail off here,
  • 6:29 - 6:31
    I figured we would have one example,
  • 6:31 - 6:32
    so, imagine you asked your robot,
  • 6:32 - 6:35
    you know, what created the moon?
  • 6:35 - 6:36
    Who better to answer it
  • 6:36 - 6:39
    than someone like
    Chris Hatfield and ISS astronauts.
  • 6:39 - 6:41
    So let's see what he has to say.
  • 6:43 - 6:45
    Robot: What created the moon?
  • 6:45 - 6:48
    Chris Hatfield: What created the moon?
  • 6:48 - 6:51
    The moon was created about
  • 6:51 - 6:52
    four billion years ago,
  • 6:52 - 6:55
    when a huge rocky early planet
  • 6:55 - 6:57
    slammed into the Earth.
  • 6:57 - 6:58
    Something about the size of Mars
  • 6:58 - 7:00
    crashed into the Earth
  • 7:00 - 7:03
    like two big wet pool balls,
  • 7:03 - 7:05
    billiard balls, slamming into each other,
  • 7:05 - 7:06
    and ripped a big section out of the Earth,
  • 7:06 - 7:08
    went off like a big globe into space
  • 7:08 - 7:10
    and started orbiting the world
  • 7:10 - 7:11
    and slowly cooled
  • 7:11 - 7:13
    and got further and further
  • 7:13 - 7:14
    from the world over the billions of years
  • 7:14 - 7:16
    until it finally coalesced
  • 7:16 - 7:18
    and looks just like our moon does today.
  • 7:18 - 7:23
    So truly, the moon is like
    the daughter of the Earth
  • 7:23 - 7:25
    (Applause)
Title:
Blab Droids -- Self-Learning Social Cloud Robots: Alex Reben at TEDxBeaconStreet
Description:

Humans are social creatures, and throughout history knowledge has been both disseminated and absorbed in a social manner, through not only written artifacts but also by oral tradition. Self-learning social cloud robots have the potential to not only be humanity's natural social window into digital knowledge, but they also afford an opportunity to collect and catalog new human knowledge in a novel way. Imagine how much uncataloged human knowledge could be saved through a natural human-centered interface which one would have feelings for like a pet: everything you teach your robot would be known by every other robot in the cloud.

more » « less
Video Language:
English
Team:
closed TED
Project:
TEDxTalks
Duration:
07:26
  • next time, please une line breaks to keep a line at 42 characters max.
    Also, note there is no space in punctuation before , . ? ! : in English
    please revise your knowledge of the use of punctuation and capitalization.

English subtitles

Revisions Compare revisions