1
00:00:00,564 --> 00:00:05,207
My job is to design, build and study
robots that communicate with people.

2
00:00:05,231 --> 00:00:08,850
But this story doesn't start with robotics
at all, it starts with animation.

3
00:00:08,874 --> 00:00:11,243
When I first saw Pixar's "Luxo Jr.,"

4
00:00:11,267 --> 00:00:13,973
I was amazed by how much
emotion they could put

5
00:00:13,997 --> 00:00:16,825
into something as trivial as a desk lamp.

6
00:00:17,477 --> 00:00:19,822
I mean, look at them --
at the end of this movie,

7
00:00:19,846 --> 00:00:22,625
you actually feel something
for two pieces of furniture.

8
00:00:22,649 --> 00:00:23,930
(Laughter)

9
00:00:23,954 --> 00:00:25,982
And I said, I have
to learn how to do this.

10
00:00:26,006 --> 00:00:28,233
So I made a really bad career decision.

11
00:00:28,257 --> 00:00:29,451
(Laughter)

12
00:00:29,475 --> 00:00:31,668
And that's what my mom
was like when I did it.

13
00:00:31,692 --> 00:00:33,833
(Laughter)

14
00:00:33,857 --> 00:00:37,278
I left a very cozy tech job in Israel
at a nice software company

15
00:00:37,302 --> 00:00:39,594
and I moved to New York
to study animation.

16
00:00:39,618 --> 00:00:40,769
And there I lived

17
00:00:40,793 --> 00:00:43,611
in a collapsing apartment building
in Harlem with roommates.

18
00:00:43,635 --> 00:00:45,659
I'm not using this phrase
metaphorically --

19
00:00:45,683 --> 00:00:48,412
the ceiling actually collapsed
one day in our living room.

20
00:00:48,436 --> 00:00:51,678
Whenever they did news stories
about building violations in New York,

21
00:00:51,702 --> 00:00:54,118
they would put the report
in front of our building,

22
00:00:54,142 --> 00:00:56,879
as kind of, like, a backdrop
to show how bad things are.

23
00:00:56,903 --> 00:00:58,840
Anyway, during the day, I went to school

24
00:00:58,864 --> 00:01:02,291
and at night I would sit and draw
frame by frame of pencil animation.

25
00:01:02,315 --> 00:01:04,500
And I learned two surprising lessons.

26
00:01:04,524 --> 00:01:09,014
One of them was that
when you want to arouse emotions,

27
00:01:09,038 --> 00:01:11,213
it doesn't matter so much
how something looks;

28
00:01:11,237 --> 00:01:14,435
it's all in the motion, in the timing
of how the thing moves.

29
00:01:14,816 --> 00:01:18,006
And the second was something
one of our teachers told us.

30
00:01:18,030 --> 00:01:20,323
He actually did the weasel in "Ice Age."

31
00:01:20,662 --> 00:01:24,519
And he said, "As an animator,
you're not a director -- you're an actor."

32
00:01:24,895 --> 00:01:27,918
So, if you want to find
the right motion for a character,

33
00:01:27,942 --> 00:01:30,387
don't think about it --
go use your body to find it.

34
00:01:30,411 --> 00:01:33,344
Stand in front of a mirror,
act it out in front of a camera --

35
00:01:33,368 --> 00:01:36,306
whatever you need -- and then
put it back in your character.

36
00:01:36,696 --> 00:01:39,908
A year later I found myself at MIT
in the Robotic Life Group.

37
00:01:39,932 --> 00:01:42,790
It was one of the first groups
researching the relationships

38
00:01:42,814 --> 00:01:44,128
between humans and robots.

39
00:01:44,152 --> 00:01:48,108
And I still had this dream
to make an actual, physical Luxo Jr. lamp.

40
00:01:48,132 --> 00:01:51,124
But I found that robots didn't move
at all in this engaging way

41
00:01:51,148 --> 00:01:53,323
that I was used to
from my animation studies.

42
00:01:53,347 --> 00:01:55,811
Instead, they were all --
how should I put it --

43
00:01:55,835 --> 00:01:57,289
they were all kind of robotic.

44
00:01:57,313 --> 00:01:59,187
(Laughter)

45
00:01:59,211 --> 00:02:02,947
And I thought, what if I took
whatever I learned in animation school,

46
00:02:02,971 --> 00:02:05,574
and used that to design
my robotic desk lamp.

47
00:02:05,598 --> 00:02:07,643
So I went and designed frame by frame

48
00:02:07,667 --> 00:02:11,627
to try to make this robot as graceful
and engaging as possible.

49
00:02:12,199 --> 00:02:15,908
And here when you see the robot
interacting with me on a desktop --

50
00:02:15,932 --> 00:02:17,968
and I'm actually redesigning the robot,

51
00:02:17,992 --> 00:02:19,796
so, unbeknownst to itself,

52
00:02:19,820 --> 00:02:22,395
it's kind of digging its own
grave by helping me.

53
00:02:22,419 --> 00:02:24,425
(Laughter)

54
00:02:24,449 --> 00:02:27,688
I wanted it to be less of a mechanical
structure giving me light,

55
00:02:27,712 --> 00:02:30,730
and more of a helpful,
kind of quiet apprentice

56
00:02:30,754 --> 00:02:33,923
that's always there when you need it
and doesn't really interfere.

57
00:02:33,947 --> 00:02:37,282
And when, for example, I'm looking
for a battery that I can't find,

58
00:02:37,306 --> 00:02:40,253
in a subtle way, it'll show me
where the battery is.

59
00:02:41,872 --> 00:02:43,705
So you can see my confusion here.

60
00:02:44,442 --> 00:02:45,654
I'm not an actor.

61
00:02:48,585 --> 00:02:51,618
And I want you to notice
how the same mechanical structure

62
00:02:51,642 --> 00:02:53,729
can, at one point,
just by the way it moves,

63
00:02:53,753 --> 00:02:55,897
seem gentle and caring
and in the other case,

64
00:02:55,921 --> 00:02:58,292
seem violent and confrontational.

65
00:02:58,316 --> 00:03:01,163
And it's the same structure,
just the motion is different.

66
00:03:07,419 --> 00:03:11,991
Actor: "You want to know something?
Well, you want to know something?

67
00:03:12,015 --> 00:03:13,904
He was already dead!

68
00:03:13,928 --> 00:03:17,830
Just laying there, eyes glazed over!"

69
00:03:17,854 --> 00:03:18,875
(Laughter)

70
00:03:18,899 --> 00:03:21,955
But, moving in a graceful way
is just one building block

71
00:03:21,979 --> 00:03:24,574
of this whole structure
called human-robot interaction.

72
00:03:24,598 --> 00:03:28,051
I was, at the time, doing my PhD,
I was working on human-robot teamwork,

73
00:03:28,075 --> 00:03:30,137
teams of humans and robots
working together.

74
00:03:30,161 --> 00:03:31,796
I was studying the engineering,

75
00:03:31,820 --> 00:03:34,352
the psychology,
the philosophy of teamwork,

76
00:03:34,376 --> 00:03:35,527
and at the same time,

77
00:03:35,551 --> 00:03:38,045
I found myself in my own kind
of teamwork situation,

78
00:03:38,069 --> 00:03:40,321
with a good friend of mine,
who's actually here.

79
00:03:40,345 --> 00:03:42,749
And in that situation,
we can easily imagine robots

80
00:03:42,773 --> 00:03:44,673
in the near future being there with us.

81
00:03:44,697 --> 00:03:46,387
It was after a Passover Seder.

82
00:03:46,411 --> 00:03:48,466
We were folding up
a lot of folding chairs,

83
00:03:48,490 --> 00:03:51,148
and I was amazed at how quickly
we found our own rhythm.

84
00:03:51,172 --> 00:03:54,282
Everybody did their own part,
we didn't have to divide our tasks.

85
00:03:54,306 --> 00:03:56,752
We didn't have to communicate
verbally about this --

86
00:03:56,776 --> 00:03:57,965
it all just happened.

87
00:03:57,989 --> 00:04:00,869
And I thought, humans and robots
don't look at all like this.

88
00:04:00,893 --> 00:04:04,024
When humans and robots interact,
it's much more like a chess game:

89
00:04:04,048 --> 00:04:07,169
the human does a thing, the robot
analyzes whatever the human did,

90
00:04:07,193 --> 00:04:09,836
the robot decides what to do next,
plans it and does it.

91
00:04:09,860 --> 00:04:12,215
Then the human waits,
until it's their turn again.

92
00:04:12,239 --> 00:04:14,970
So it's much more like a chess game,
and that makes sense,

93
00:04:14,994 --> 00:04:18,105
because chess is great for mathematicians
and computer scientists.

94
00:04:18,129 --> 00:04:21,557
It's all about information, analysis,
decision-making and planning.

95
00:04:21,581 --> 00:04:25,311
But I wanted my robot
to be less of a chess player,

96
00:04:25,335 --> 00:04:27,128
and more like a doer

97
00:04:27,152 --> 00:04:29,153
that just clicks and works together.

98
00:04:29,177 --> 00:04:32,546
So I made my second
horrible career choice:

99
00:04:32,570 --> 00:04:35,123
I decided to study acting for a semester.

100
00:04:35,147 --> 00:04:38,066
I took off from the PhD,
I went to acting classes.

101
00:04:38,090 --> 00:04:40,128
I actually participated in a play --

102
00:04:40,152 --> 00:04:42,454
I hope there’s no video
of that around still.

103
00:04:42,478 --> 00:04:43,516
(Laughter)

104
00:04:43,540 --> 00:04:45,834
And I got every book
I could find about acting,

105
00:04:45,858 --> 00:04:48,987
including one from the 19th century
that I got from the library.

106
00:04:49,011 --> 00:04:52,624
And I was really amazed, because my name
was the second name on the list --

107
00:04:52,648 --> 00:04:54,648
the previous name was in 1889.

108
00:04:54,672 --> 00:04:55,687
(Laughter)

109
00:04:55,711 --> 00:04:57,960
And this book was
kind of waiting for 100 years

110
00:04:57,984 --> 00:05:00,316
to be rediscovered for robotics.

111
00:05:00,340 --> 00:05:01,916
And this book shows actors

112
00:05:01,940 --> 00:05:04,178
how to move every muscle in the body

113
00:05:04,202 --> 00:05:06,918
to match every kind of emotion
that they want to express.

114
00:05:06,942 --> 00:05:09,968
But the real revelation was
when I learned about method acting.

115
00:05:09,992 --> 00:05:12,245
It became very popular
in the 20th century.

116
00:05:12,269 --> 00:05:13,420
And method acting said

117
00:05:13,444 --> 00:05:15,744
you don't have to plan
every muscle in your body;

118
00:05:15,768 --> 00:05:18,713
instead, you have to use your body
to find the right movement.

119
00:05:18,737 --> 00:05:21,617
You have to use your sense memory
to reconstruct the emotions

120
00:05:21,641 --> 00:05:24,679
and kind of think with your body
to find the right expression --

121
00:05:24,703 --> 00:05:26,578
improvise, play off your scene partner.

122
00:05:26,602 --> 00:05:28,047
And this came at the same time

123
00:05:28,071 --> 00:05:31,177
as I was reading about this trend
in cognitive psychology,

124
00:05:31,201 --> 00:05:34,453
called embodied cognition,
which also talks about the same ideas.

125
00:05:34,477 --> 00:05:35,992
We use our bodies to think;

126
00:05:36,016 --> 00:05:38,973
we don't just think with our brains
and use our bodies to move,

127
00:05:38,997 --> 00:05:40,915
but our bodies feed back into our brain

128
00:05:40,939 --> 00:05:43,104
to generate the way that we behave.

129
00:05:43,128 --> 00:05:44,724
And it was like a lightning bolt.

130
00:05:44,748 --> 00:05:45,962
I went back to my office,

131
00:05:45,986 --> 00:05:48,391
I wrote this paper,
which I never really published,

132
00:05:48,415 --> 00:05:50,947
called "Acting Lessons
for Artificial Intelligence."

133
00:05:50,971 --> 00:05:52,352
And I even took another month

134
00:05:52,376 --> 00:05:54,820
to do what was then the first theater play

135
00:05:54,844 --> 00:05:56,797
with a human and a robot acting together.

136
00:05:56,821 --> 00:05:59,214
That's what you saw
before with the actors.

137
00:06:00,564 --> 00:06:01,715
And I thought:

138
00:06:01,739 --> 00:06:04,676
How can we make an artificial
intelligence model --

139
00:06:04,700 --> 00:06:06,683
a computer, computational model --

140
00:06:06,707 --> 00:06:09,214
that will model some of these
ideas of improvisation,

141
00:06:09,238 --> 00:06:11,068
of taking risks, of taking chances,

142
00:06:11,092 --> 00:06:12,619
even of making mistakes?

143
00:06:12,643 --> 00:06:15,256
Maybe it can make for better
robotic teammates.

144
00:06:15,280 --> 00:06:17,884
So I worked for quite
a long time on these models

145
00:06:17,908 --> 00:06:20,303
and I implemented them
on a number of robots.

146
00:06:20,327 --> 00:06:22,628
Here you can see a very early example

147
00:06:22,652 --> 00:06:26,231
with the robots trying to use
this embodied artificial intelligence

148
00:06:26,255 --> 00:06:28,718
to try to match my movements
as closely as possible.

149
00:06:28,742 --> 00:06:30,205
It's sort of like a game.

150
00:06:30,530 --> 00:06:31,739
Let's look at it.

151
00:06:35,652 --> 00:06:39,191
You can see when I psych
it out, it gets fooled.

152
00:06:39,698 --> 00:06:42,293
And it's a little bit
like what you might see actors do

153
00:06:42,317 --> 00:06:43,951
when they try to mirror each other

154
00:06:43,975 --> 00:06:46,340
to find the right synchrony between them.

155
00:06:46,364 --> 00:06:48,196
And then, I did another experiment,

156
00:06:48,220 --> 00:06:52,279
and I got people off the street
to use the robotic desk lamp,

157
00:06:52,303 --> 00:06:55,405
and try out this idea
of embodied artificial intelligence.

158
00:06:55,921 --> 00:07:00,519
So, I actually used two kinds
of brains for the same robot.

159
00:07:00,543 --> 00:07:02,495
The robot is the same lamp that you saw,

160
00:07:02,519 --> 00:07:03,858
and I put two brains in it.

161
00:07:03,882 --> 00:07:05,609
For one half of the people,

162
00:07:05,633 --> 00:07:08,569
I put in a brain
that's kind of the traditional,

163
00:07:08,593 --> 00:07:09,822
calculated robotic brain.

164
00:07:09,846 --> 00:07:12,486
It waits for its turn,
it analyzes everything, it plans.

165
00:07:12,510 --> 00:07:14,192
Let's call it the calculated brain.

166
00:07:14,216 --> 00:07:17,463
The other got more the stage
actor, risk-taker brain.

167
00:07:17,487 --> 00:07:19,609
Let's call it the adventurous brain.

168
00:07:19,633 --> 00:07:22,570
It sometimes acts without knowing
everything it has to know.

169
00:07:22,594 --> 00:07:24,886
It sometimes makes mistakes
and corrects them.

170
00:07:24,910 --> 00:07:28,828
And I had them do this very tedious task
that took almost 20 minutes,

171
00:07:28,852 --> 00:07:30,328
and they had to work together,

172
00:07:30,352 --> 00:07:32,905
somehow simulating, like, a factory job

173
00:07:32,929 --> 00:07:35,016
of repetitively doing the same thing.

174
00:07:35,445 --> 00:07:38,776
What I found is that people
actually loved the adventurous robot.

175
00:07:38,800 --> 00:07:40,577
They thought it was more intelligent,

176
00:07:40,601 --> 00:07:42,707
more committed,
a better member of the team,

177
00:07:42,731 --> 00:07:44,796
contributed to the success
of the team more.

178
00:07:44,820 --> 00:07:46,528
They even called it "he" and "she,"

179
00:07:46,552 --> 00:07:48,673
whereas people with the calculated brain

180
00:07:48,697 --> 00:07:51,499
called it "it," and nobody
ever called it "he" or "she."

181
00:07:51,876 --> 00:07:55,169
When they talked about it after the task,
with the adventurous brain,

182
00:07:55,193 --> 00:07:58,976
they said, "By the end, we were good
friends and high-fived mentally."

183
00:07:59,397 --> 00:08:00,690
Whatever that means.

184
00:08:00,714 --> 00:08:02,620
(Laughter)

185
00:08:02,644 --> 00:08:03,997
Sounds painful.

186
00:08:04,021 --> 00:08:06,759
Whereas the people
with the calculated brain

187
00:08:06,783 --> 00:08:09,141
said it was just like a lazy apprentice.

188
00:08:09,165 --> 00:08:11,910
It only did what it was supposed
to do and nothing more,

189
00:08:11,934 --> 00:08:14,283
which is almost what people
expect robots to do,

190
00:08:14,307 --> 00:08:17,862
so I was surprised that people
had higher expectations of robots

191
00:08:17,886 --> 00:08:21,392
than what anybody in robotics
thought robots should be doing.

192
00:08:22,027 --> 00:08:24,042
And in a way, I thought,
maybe it's time --

193
00:08:24,066 --> 00:08:27,028
just like method acting
changed the way people thought

194
00:08:27,052 --> 00:08:28,648
about acting in the 19th century,

195
00:08:28,672 --> 00:08:31,749
from going from the very calculated,
planned way of behaving,

196
00:08:31,773 --> 00:08:35,271
to a more intuitive, risk-taking,
embodied way of behaving --

197
00:08:35,295 --> 00:08:38,491
maybe it's time for robots
to have the same kind of revolution.

198
00:08:39,994 --> 00:08:43,593
A few years later, I was at my next
research job at Georgia Tech in Atlanta,

199
00:08:43,617 --> 00:08:46,441
and I was working in a group
dealing with robotic musicians.

200
00:08:46,465 --> 00:08:48,825
And I thought, music:
that's the perfect place

201
00:08:48,849 --> 00:08:52,992
to look at teamwork, coordination,
timing, improvisation --

202
00:08:53,016 --> 00:08:55,459
and we just got this robot
playing marimba.

203
00:08:55,483 --> 00:08:57,739
And the marimba, for everybody like me,

204
00:08:57,763 --> 00:09:00,502
it was this huge, wooden xylophone.

205
00:09:00,526 --> 00:09:02,589
And when I was looking at this,

206
00:09:02,613 --> 00:09:05,611
I looked at other works
in human-robot improvisation --

207
00:09:05,635 --> 00:09:08,382
yes, there are other works
in human-robot improvisation --

208
00:09:08,406 --> 00:09:10,755
and they were also a little bit
like a chess game.

209
00:09:10,779 --> 00:09:11,930
The human would play,

210
00:09:11,954 --> 00:09:14,127
the robot analyzed what was played,

211
00:09:14,151 --> 00:09:16,230
and would improvise their own part.

212
00:09:16,254 --> 00:09:19,392
So, this is what musicians called
a call-and-response interaction,

213
00:09:19,416 --> 00:09:23,163
and it also fits very well
robots and artificial intelligence.

214
00:09:23,187 --> 00:09:26,440
But I thought, if I use the same ideas
I used in the theater play

215
00:09:26,464 --> 00:09:28,208
and in the teamwork studies,

216
00:09:28,232 --> 00:09:31,882
maybe I can make the robots
jam together like a band.

217
00:09:31,906 --> 00:09:36,174
Everybody's riffing off each other,
nobody is stopping for a moment.

218
00:09:36,198 --> 00:09:39,072
And so I tried to do the same
things, this time with music,

219
00:09:39,096 --> 00:09:41,930
where the robot doesn't really know
what it's about to play,

220
00:09:41,954 --> 00:09:44,971
it just sort of moves its body
and uses opportunities to play,

221
00:09:44,995 --> 00:09:47,534
and does what my jazz teacher
when I was 17 taught me.

222
00:09:47,558 --> 00:09:48,955
She said, when you improvise,

223
00:09:48,979 --> 00:09:51,995
sometimes you don't know
what you're doing, and you still do it.

224
00:09:52,019 --> 00:09:55,345
So I tried to make a robot that doesn't
actually know what it's doing,

225
00:09:55,369 --> 00:09:56,590
but is still doing it.

226
00:09:56,614 --> 00:09:59,125
So let's look at a few seconds
from this performance,

227
00:09:59,149 --> 00:10:01,276
where the robot listens
to the human musician

228
00:10:01,300 --> 00:10:02,465
and improvises.

229
00:10:02,489 --> 00:10:05,464
And then, look how the human
musician also responds

230
00:10:05,488 --> 00:10:06,761
to what the robot is doing

231
00:10:06,785 --> 00:10:09,429
and picking up from its behavior,

232
00:10:09,453 --> 00:10:13,409
and at some point can even be surprised
by what the robot came up with.

233
00:10:13,433 --> 00:10:15,627
(Music)

234
00:10:55,618 --> 00:10:57,618
(Music ends)

235
00:10:59,134 --> 00:11:04,868
(Applause)

236
00:11:04,892 --> 00:11:07,142
Being a musician
is not just about making notes,

237
00:11:07,166 --> 00:11:09,514
otherwise nobody
would ever go see a live show.

238
00:11:09,538 --> 00:11:11,660
Musicians also communicate
with their bodies,

239
00:11:11,684 --> 00:11:13,701
with other band members,
with the audience,

240
00:11:13,725 --> 00:11:15,739
they use their bodies
to express the music.

241
00:11:15,763 --> 00:11:18,460
And I thought, we already have
a robot musician on stage,

242
00:11:18,484 --> 00:11:20,796
why not make it be
a full-fledged musician?

243
00:11:20,820 --> 00:11:23,494
And I started designing
a socially expressive head

244
00:11:23,518 --> 00:11:24,861
for the robot.

245
00:11:24,885 --> 00:11:26,955
The head doesn’t actually
touch the marimba,

246
00:11:26,979 --> 00:11:28,938
it just expresses what the music is like.

247
00:11:28,962 --> 00:11:31,458
These are some napkin sketches
from a bar in Atlanta

248
00:11:31,482 --> 00:11:34,025
that was dangerously
located exactly halfway

249
00:11:34,049 --> 00:11:35,860
between my lab and my home.

250
00:11:35,884 --> 00:11:39,822
So I spent, I would say, on average,
three to four hours a day there.

251
00:11:39,846 --> 00:11:41,033
I think.

252
00:11:41,057 --> 00:11:42,584
(Laughter)

253
00:11:42,917 --> 00:11:45,799
And I went back to my animation
tools and tried to figure out

254
00:11:45,823 --> 00:11:48,193
not just what a robotic
musician would look like,

255
00:11:48,217 --> 00:11:50,842
but especially what a robotic
musician would move like,

256
00:11:50,866 --> 00:11:54,268
to sort of show that it doesn't like
what the other person is playing --

257
00:11:54,292 --> 00:11:58,117
and maybe show whatever beat
it's feeling at the moment.

258
00:11:58,141 --> 00:12:02,644
So we ended up actually getting the money
to build this robot, which was nice.

259
00:12:02,668 --> 00:12:05,272
I'm going to show you now
the same kind of performance,

260
00:12:05,296 --> 00:12:07,336
this time with a socially expressive head.

261
00:12:07,360 --> 00:12:09,160
And notice one thing --

262
00:12:09,184 --> 00:12:10,855
how the robot is really showing us

263
00:12:10,879 --> 00:12:12,785
the beat it's picking up from the human,

264
00:12:12,809 --> 00:12:16,736
while also giving the human a sense
that the robot knows what it's doing.

265
00:12:16,760 --> 00:12:18,680
And also how it changes the way it moves

266
00:12:18,704 --> 00:12:20,656
as soon as it starts its own solo.

267
00:12:20,680 --> 00:12:24,576
(Music)

268
00:12:24,600 --> 00:12:27,620
Now it's looking at me,
showing that it's listening.

269
00:12:27,644 --> 00:12:30,367
(Music)

270
00:12:49,181 --> 00:12:52,102
Now look at the final chord
of the piece again.

271
00:12:52,126 --> 00:12:55,043
And this time the robot
communicates with its body

272
00:12:55,067 --> 00:12:57,180
when it's busy doing its own thing,

273
00:12:57,204 --> 00:13:02,055
and when it's ready to coordinate
the final chord with me.

274
00:13:02,079 --> 00:13:04,816
(Music)

275
00:13:09,783 --> 00:13:11,783
(Music ending)

276
00:13:14,103 --> 00:13:15,104
(Final chord)

277
00:13:15,128 --> 00:13:20,882
(Applause)

278
00:13:20,906 --> 00:13:22,061
Thanks.

279
00:13:22,085 --> 00:13:23,666
I hope you see

280
00:13:23,690 --> 00:13:28,086
how much this part of the body
that doesn't touch the instrument

281
00:13:28,110 --> 00:13:30,702
actually helps
with the musical performance.

282
00:13:31,160 --> 00:13:33,169
And at some point -- we are in Atlanta,

283
00:13:33,193 --> 00:13:36,443
so obviously some rapper
will come into our lab at some point --

284
00:13:36,467 --> 00:13:41,260
and we had this rapper come in and do
a little jam with the robot.

285
00:13:41,284 --> 00:13:45,332
Here you can see the robot
basically responding to the beat.

286
00:13:45,356 --> 00:13:46,507
Notice two things:

287
00:13:46,531 --> 00:13:50,140
one, how irresistible it is to join
the robot while it's moving its head.

288
00:13:50,164 --> 00:13:52,767
You kind of want to move
your own head when it does it.

289
00:13:52,791 --> 00:13:56,377
And second, even though the rapper
is really focused on his iPhone,

290
00:13:56,401 --> 00:13:59,511
as soon as the robot turns
to him, he turns back.

291
00:13:59,535 --> 00:14:02,171
So even though it's just
in the periphery of his vision,

292
00:14:02,195 --> 00:14:04,315
in the corner of his eye,
it's very powerful.

293
00:14:04,339 --> 00:14:06,173
And the reason is that we can't ignore

294
00:14:06,197 --> 00:14:08,221
physical things moving in our environment.

295
00:14:08,245 --> 00:14:09,415
We are wired for that.

296
00:14:09,439 --> 00:14:11,083
So if you have a problem --

297
00:14:11,107 --> 00:14:15,782
maybe your partner is looking
at their iPhone or smartphone too much --

298
00:14:15,806 --> 00:14:18,640
you might want to have a robot
there to get their attention.

299
00:14:18,664 --> 00:14:19,665
(Laughter)

300
00:14:19,689 --> 00:14:21,758
(Music)

301
00:14:34,113 --> 00:14:36,113
(Music ends)

302
00:14:38,138 --> 00:14:45,086
(Applause)

303
00:14:45,633 --> 00:14:49,467
Just to introduce the last robot
that we've worked on,

304
00:14:49,491 --> 00:14:51,841
it came out of something
surprising that we found:

305
00:14:51,865 --> 00:14:55,023
Some point people didn't care
about the robot being intelligent,

306
00:14:55,047 --> 00:14:56,449
able to improvise and listen,

307
00:14:56,473 --> 00:15:00,879
and do all these embodied intelligence
things that I spent years developing.

308
00:15:00,903 --> 00:15:03,584
They really liked that the robot
was enjoying the music.

309
00:15:03,608 --> 00:15:04,609
(Laughter)

310
00:15:04,633 --> 00:15:07,178
And they didn't say the robot
was moving to the music,

311
00:15:07,202 --> 00:15:08,690
they said "enjoying" the music.

312
00:15:08,714 --> 00:15:10,840
And we thought,
why don't we take this idea,

313
00:15:10,864 --> 00:15:13,605
and I designed a new piece of furniture.

314
00:15:13,629 --> 00:15:16,201
This time it wasn't a desk lamp,
it was a speaker dock,

315
00:15:16,225 --> 00:15:18,865
one of those things
you plug your smartphone in.

316
00:15:18,889 --> 00:15:20,039
And I thought,

317
00:15:20,063 --> 00:15:23,563
what would happen if your speaker dock
didn't just play the music for you,

318
00:15:23,587 --> 00:15:25,633
but would actually enjoy it, too?

319
00:15:25,657 --> 00:15:29,818
And so again, here are some
animation tests from an early stage.

320
00:15:29,842 --> 00:15:31,226
(Laughter)

321
00:15:31,835 --> 00:15:34,706
And this is what the final
product looked like.

322
00:15:46,247 --> 00:15:48,428
(Music)

323
00:16:06,848 --> 00:16:08,848
(Music ends)

324
00:16:09,461 --> 00:16:12,081
So, a lot of bobbing heads.

325
00:16:12,105 --> 00:16:15,716
(Applause)

326
00:16:15,740 --> 00:16:17,630
A lot of bobbing heads in the audience,

327
00:16:17,654 --> 00:16:20,362
so we can still see
robots influence people.

328
00:16:20,386 --> 00:16:23,171
And it's not just fun and games.

329
00:16:23,195 --> 00:16:25,396
I think one of the reasons I care so much

330
00:16:25,420 --> 00:16:27,653
about robots that use
their body to communicate

331
00:16:27,677 --> 00:16:29,383
and use their body to move is --

332
00:16:29,407 --> 00:16:32,797
I'm going to let you in on a little
secret we roboticists are hiding --

333
00:16:32,821 --> 00:16:35,614
is that every one of you
is going to be living with a robot

334
00:16:35,638 --> 00:16:37,230
at some point in your life.

335
00:16:37,254 --> 00:16:40,187
Somewhere in your future,
there will be a robot in your life.

336
00:16:40,211 --> 00:16:42,079
If not in yours, your children's lives.

337
00:16:42,103 --> 00:16:47,093
And I want these robots to be more fluent,
more engaging, more graceful

338
00:16:47,117 --> 00:16:48,846
than currently they seem to be.

339
00:16:48,870 --> 00:16:52,157
And for that I think maybe robots
need to be less like chess players

340
00:16:52,181 --> 00:16:54,822
and more like stage actors
and more like musicians.

341
00:16:54,846 --> 00:16:57,659
Maybe they should be able
to take chances and improvise.

342
00:16:57,683 --> 00:17:00,707
Maybe they should be able
to anticipate what you're about to do.

343
00:17:00,731 --> 00:17:04,079
Maybe they even need to be able
to make mistakes and correct them,

344
00:17:04,103 --> 00:17:05,991
because in the end, we are human.

345
00:17:06,015 --> 00:17:09,569
And maybe as humans,
robots that are a little less than perfect

346
00:17:09,593 --> 00:17:11,324
are just perfect for us.

347
00:17:11,348 --> 00:17:12,507
Thank you.

348
00:17:12,531 --> 00:17:19,367
(Applause)