1
00:00:01,333 --> 00:00:02,809
Food delivery.

2
00:00:02,833 --> 00:00:05,809
It's the thing that saves
millennials from starvation.

3
00:00:06,770 --> 00:00:08,188
By my calculations,

4
00:00:08,212 --> 00:00:11,950
Americans order over 20 million
restaurant deliveries

5
00:00:11,974 --> 00:00:13,640
every single day.

6
00:00:14,022 --> 00:00:17,458
Over half of these deliveries
are actually within walking distance.

7
00:00:17,482 --> 00:00:21,233
But nine out of 10 are delivered in cars.

8
00:00:21,607 --> 00:00:23,933
So basically,

9
00:00:23,957 --> 00:00:25,861
we are moving a two-pound burrito

10
00:00:25,885 --> 00:00:27,655
in a two-ton car

11
00:00:27,679 --> 00:00:29,914
20 million times a day.

12
00:00:29,938 --> 00:00:32,072
The energy to bring you that burrito

13
00:00:32,096 --> 00:00:34,850
is actually bringing you
a two-ton metal cage

14
00:00:34,874 --> 00:00:36,263
with heated seats.

15
00:00:36,287 --> 00:00:37,609
Let's be honest.

16
00:00:37,633 --> 00:00:39,633
We are addicted to our cars.

17
00:00:40,085 --> 00:00:41,935
Did you know that in America,

18
00:00:41,959 --> 00:00:45,434
for every car,
there are four parking spots?

19
00:00:45,458 --> 00:00:46,831
In some downtowns,

20
00:00:46,855 --> 00:00:49,942
over half of the real estate is for cars.

21
00:00:50,427 --> 00:00:53,401
We have designed our cities
around our cars,

22
00:00:53,425 --> 00:00:56,290
because we drive
whether we're going two miles

23
00:00:56,314 --> 00:00:57,861
or 200 miles.

24
00:00:57,885 --> 00:01:00,290
Solo, or with our whole family.

25
00:01:00,688 --> 00:01:05,600
We get into the same SUV
to go buy coffee or a coffee table.

26
00:01:06,019 --> 00:01:08,796
If we could free up
some of these streets and parking lots,

27
00:01:08,820 --> 00:01:10,201
we could build more housing,

28
00:01:10,225 --> 00:01:12,328
more social spaces, more parks.

29
00:01:12,661 --> 00:01:14,014
But to do that,

30
00:01:14,038 --> 00:01:17,942
first, we need to rethink
how we are using cars today.

31
00:01:17,966 --> 00:01:19,395
In the city of the future,

32
00:01:19,419 --> 00:01:23,065
if you want to go five blocks,
you summon a bike or a scooter.

33
00:01:23,089 --> 00:01:26,376
If you're in a rush,
a passenger drone would pick you up.

34
00:01:26,400 --> 00:01:29,661
And if you need food,
no need to have someone drive over --

35
00:01:29,685 --> 00:01:32,271
the food will make its way to you.

36
00:01:32,295 --> 00:01:35,959
Let's go back to those 20 million a day
restaurant deliveries.

37
00:01:36,341 --> 00:01:39,150
If we could get
these deliveries off the road,

38
00:01:39,174 --> 00:01:40,609
we could reduce the need

39
00:01:40,633 --> 00:01:45,363
for as many as one and a half
million cars just in the US.

40
00:01:45,387 --> 00:01:47,833
That's twice the size of San Francisco.

41
00:01:48,171 --> 00:01:52,091
Now, think of the impact this could have
on cities like Delhi,

42
00:01:52,115 --> 00:01:54,545
or my birth city of Tehran,

43
00:01:54,569 --> 00:01:58,053
where car pollution is killing
thousands of people every year.

44
00:01:58,546 --> 00:02:03,455
So how do we get some
of these deliveries off the road?

45
00:02:04,661 --> 00:02:06,391
Well, that's the question

46
00:02:06,415 --> 00:02:09,913
that my team and I have been obsessed with
over the last three years.

47
00:02:10,380 --> 00:02:13,475
And the solution is
actually one of the building blocks

48
00:02:13,499 --> 00:02:15,341
of the city of the future.

49
00:02:15,944 --> 00:02:20,155
We've been creating
small, self-driving robots

50
00:02:20,179 --> 00:02:22,806
that navigate quiet alleys and sidewalks

51
00:02:22,830 --> 00:02:24,512
on a walking pace

52
00:02:24,536 --> 00:02:28,106
and have a secured cargo
to deliver you food and supplies.

53
00:02:28,130 --> 00:02:31,304
Now, before I tell you
more about the robots,

54
00:02:31,328 --> 00:02:33,328
let's do a quick thought experiment.

55
00:02:33,947 --> 00:02:38,372
In your mind, picture a city
with thousands of robots.

56
00:02:39,007 --> 00:02:40,341
Is it this one?

57
00:02:40,761 --> 00:02:44,601
This Hollywood dystopia
is what a lot of people expect.

58
00:02:44,625 --> 00:02:50,307
But our job is to create a friendly future
that's designed for people.

59
00:02:50,331 --> 00:02:53,298
So instead of making aliens,

60
00:02:53,322 --> 00:02:57,004
we set out to create robots
that are familiar.

61
00:02:57,361 --> 00:03:00,100
Robots that would belong
in our communities.

62
00:03:00,665 --> 00:03:02,720
But we also wanted a little surprise.

63
00:03:02,744 --> 00:03:05,505
Something unexpectedly delightful.

64
00:03:05,815 --> 00:03:06,974
Think about it.

65
00:03:06,998 --> 00:03:08,481
You're walking down the street,

66
00:03:08,505 --> 00:03:10,710
and you see your very first robot.

67
00:03:10,734 --> 00:03:13,052
That's the moment
when you're going to decide

68
00:03:13,076 --> 00:03:15,861
if this is a future you love or fear.

69
00:03:16,861 --> 00:03:20,137
And with a lot of people
having these dystopian ideas,

70
00:03:20,161 --> 00:03:22,383
we need to open their minds.

71
00:03:22,407 --> 00:03:24,772
We want to surprise and delight them,

72
00:03:24,796 --> 00:03:27,952
so that we can win them over
on first impression.

73
00:03:27,976 --> 00:03:29,380
This is what we came up with.

74
00:03:29,404 --> 00:03:32,174
It's familiar, but it's also surprising.

75
00:03:32,555 --> 00:03:34,483
It's just a shopping cart,

76
00:03:34,507 --> 00:03:38,267
but it also looks like
we crossbred WALL-E with Minions.

77
00:03:39,004 --> 00:03:41,853
If you live in San Francisco
or Los Angeles,

78
00:03:41,877 --> 00:03:44,933
chances are one of these
has already delivered your food.

79
00:03:45,482 --> 00:03:48,244
As soon as we put robots
out on the street,

80
00:03:48,268 --> 00:03:50,415
we learned some
really interesting problems.

81
00:03:50,439 --> 00:03:53,400
Like, how should robots cross the road?

82
00:03:53,704 --> 00:03:59,330
Or how should robots interact with people
with visual or mobility impairments?

83
00:04:00,149 --> 00:04:03,805
We quickly realized
that we need to teach our robots

84
00:04:03,829 --> 00:04:06,099
how to communicate with people.

85
00:04:06,877 --> 00:04:09,718
People on the sidewalk
come from every walk of life,

86
00:04:09,742 --> 00:04:12,960
so we needed to create a new language,

87
00:04:12,984 --> 00:04:14,729
kind of a universal language

88
00:04:14,753 --> 00:04:17,864
so people and robots
can understand each other

89
00:04:17,888 --> 00:04:19,253
right off the bat.

90
00:04:19,277 --> 00:04:22,345
Because no one is going
to be reading user manuals.

91
00:04:23,719 --> 00:04:27,441
We started with eyes,
because eyes are universal.

92
00:04:27,465 --> 00:04:29,403
They can show where the robot is going

93
00:04:29,427 --> 00:04:31,513
or if it's confused.

94
00:04:31,537 --> 00:04:34,378
Plus, eyes make robots more human.

95
00:04:34,871 --> 00:04:37,045
We also used sounds.

96
00:04:37,069 --> 00:04:39,534
For example, we created this running sound

97
00:04:39,558 --> 00:04:41,379
with frequent gaps

98
00:04:41,403 --> 00:04:45,268
so that people with visual impairments
could locate their robots

99
00:04:45,292 --> 00:04:47,244
using the Doppler effect.

100
00:04:47,268 --> 00:04:50,061
But it turned out these were not enough.

101
00:04:50,085 --> 00:04:51,299
At intersections,

102
00:04:51,323 --> 00:04:53,323
cars would cut in front of our robots.

103
00:04:53,347 --> 00:04:55,863
Drivers were getting confused sometimes,

104
00:04:55,887 --> 00:04:59,931
because robots would take too long
before they started crossing.

105
00:05:00,630 --> 00:05:03,467
Even ordinary pedestrians
were getting confused.

106
00:05:03,796 --> 00:05:07,188
Sometimes, they couldn't figure out
on which side to pass the robots,

107
00:05:07,212 --> 00:05:10,640
because robots make a lot
of small adjustments to their direction

108
00:05:10,664 --> 00:05:11,836
as they move.

109
00:05:11,860 --> 00:05:14,312
This actually sparked a new idea.

110
00:05:14,625 --> 00:05:18,669
What if we used movement
to create a universal language?

111
00:05:19,106 --> 00:05:20,772
Like, at intersections,

112
00:05:20,796 --> 00:05:24,201
robots would gently move forward
before they start crossing,

113
00:05:24,225 --> 00:05:26,820
to signal to drivers that it's their turn.

114
00:05:27,928 --> 00:05:29,864
If they see someone in a wheelchair,

115
00:05:29,888 --> 00:05:33,982
they yield by pointing themselves
away from the sidewalk,

116
00:05:34,006 --> 00:05:36,299
to signal that they're not going to move.

117
00:05:37,704 --> 00:05:39,371
Some of you may remember this.

118
00:05:39,395 --> 00:05:44,833
In 2015, Canadian researchers
sent a robot hitchhiking across the US.

119
00:05:45,677 --> 00:05:47,304
It didn't get very far.

120
00:05:47,328 --> 00:05:51,360
It turns out that robots
can also use some social skills.

121
00:05:51,384 --> 00:05:53,498
Like, if they're being tampered with,

122
00:05:53,522 --> 00:05:58,934
Carnegie Mellon researchers have shown
that small toy robots should play dead,

123
00:05:58,958 --> 00:06:01,800
because people feel bad
when they think they broke it.

124
00:06:02,474 --> 00:06:04,490
But delivery robots aren't toys,

125
00:06:04,514 --> 00:06:07,069
they're not small,
they are out there in public.

126
00:06:07,394 --> 00:06:10,371
We found that with delivery robots,

127
00:06:10,395 --> 00:06:12,331
to get people to stop tampering,

128
00:06:12,355 --> 00:06:14,355
robots need to show awareness.

129
00:06:14,950 --> 00:06:17,410
It's kind of the opposite of playing dead.

130
00:06:17,792 --> 00:06:18,991
In this case,

131
00:06:19,015 --> 00:06:21,656
robots need to acknowledge the situation

132
00:06:21,680 --> 00:06:23,680
to get people to step away.

133
00:06:24,546 --> 00:06:25,991
Also, a word of advice.

134
00:06:26,015 --> 00:06:28,910
If you are a robot and you see small kids,

135
00:06:28,934 --> 00:06:30,934
run towards the closest adult.

136
00:06:31,315 --> 00:06:35,066
It turns out that some kids
just love harassing robots.

137
00:06:35,455 --> 00:06:38,010
So besides dystopia,

138
00:06:38,034 --> 00:06:41,098
Hollywood also promised us
some really cool robots

139
00:06:41,122 --> 00:06:43,983
that would run our errands
or keep us company.

140
00:06:44,007 --> 00:06:46,880
So far, we've really focused
on food delivery,

141
00:06:46,904 --> 00:06:48,555
but in the future,

142
00:06:48,579 --> 00:06:50,537
these robots can do more.

143
00:06:50,561 --> 00:06:55,252
Like, they could gather excess food
and bring it to shelters every night.

144
00:06:55,593 --> 00:06:59,593
Because in America,
we waste 30 percent of our food,

145
00:06:59,617 --> 00:07:03,145
while 10 percent of our people
experience food insecurity.

146
00:07:03,169 --> 00:07:05,169
These robots could be
part of the solution.

147
00:07:06,003 --> 00:07:10,018
Or when we have hundreds of robots
running around cities,

148
00:07:10,042 --> 00:07:14,285
we could have robots carry
emergency medications at all times,

149
00:07:14,309 --> 00:07:17,539
just in case someone nearby
has an allergic reaction

150
00:07:17,563 --> 00:07:19,071
or an asthma attack.

151
00:07:19,095 --> 00:07:22,291
These robots could be on-site
within a minute or two,

152
00:07:22,315 --> 00:07:23,838
faster than anyone else.

153
00:07:23,862 --> 00:07:25,498
And during pandemics,

154
00:07:25,522 --> 00:07:28,005
robots can be a key part
of our infrastructure.

155
00:07:28,458 --> 00:07:29,807
They can ensure

156
00:07:29,831 --> 00:07:33,418
that we can provide our communities
with the essential needs

157
00:07:33,442 --> 00:07:35,069
even during emergencies.

158
00:07:35,093 --> 00:07:37,616
Let me leave you with one last thought.

159
00:07:38,331 --> 00:07:43,631
Today, objects can't get from A to B
without human help,

160
00:07:43,655 --> 00:07:47,329
because our three-dimensional world
is quite complex.

161
00:07:47,814 --> 00:07:51,532
But new sensors and AI can change that.

162
00:07:52,120 --> 00:07:54,826
In a way, technology is like a baby

163
00:07:54,850 --> 00:07:59,260
that has just learned to recognize objects
and understand words,

164
00:07:59,284 --> 00:08:02,156
and maybe even hold a basic conversation,

165
00:08:02,180 --> 00:08:04,180
but it hasn't learned to walk yet.

166
00:08:04,934 --> 00:08:07,498
Now, we are teaching technology

167
00:08:07,522 --> 00:08:09,806
how to navigate
the three-dimensional world

168
00:08:09,830 --> 00:08:11,292
without our help.

169
00:08:12,585 --> 00:08:14,664
We are entering this new era

170
00:08:14,688 --> 00:08:19,315
where insentient objects
are going to get up and move freely.

171
00:08:20,305 --> 00:08:21,791
And when they do,

172
00:08:21,815 --> 00:08:24,473
we've got to make sure
they don't look like aliens.

173
00:08:24,497 --> 00:08:29,236
My vision for the future
is that when things come to life,

174
00:08:29,260 --> 00:08:30,720
they do so with joy.

175
00:08:31,053 --> 00:08:34,095
You know, less like
the movie "Terminator"

176
00:08:34,119 --> 00:08:35,786
and more like "Toy Story."

177
00:08:36,159 --> 00:08:37,373
Thank you.