1
00:00:01,747 --> 00:00:03,234
This is Henry,

2
00:00:03,258 --> 00:00:05,064
a cute boy,

3
00:00:05,088 --> 00:00:07,226
and when Henry was three,

4
00:00:07,250 --> 00:00:12,185
his mom found him having
some febrile seizures.

5
00:00:13,080 --> 00:00:18,040
Febrile seizures are seizures that occur
when you also have a fever,

6
00:00:18,064 --> 00:00:19,962
and the doctor said,

7
00:00:19,986 --> 00:00:22,728
"Don't worry too much.
Kids usually outgrow these."

8
00:00:23,506 --> 00:00:27,487
When he was four,
he had a convulsive seizure,

9
00:00:27,511 --> 00:00:30,773
the kind that you lose
consciousness and shake --

10
00:00:30,797 --> 00:00:33,693
a generalized tonic-clonic seizure --

11
00:00:33,717 --> 00:00:40,701
and while the diagnosis of epilepsy
was in the mail,

12
00:00:41,763 --> 00:00:44,439
Henry's mom went to get him
out of bed one morning,

13
00:00:45,324 --> 00:00:47,228
and as she went in his room,

14
00:00:48,753 --> 00:00:53,443
she found his cold, lifeless body.

15
00:00:55,904 --> 00:00:57,960
Henry died of SUDEP,

16
00:00:57,984 --> 00:01:00,389
sudden unexpected death in epilepsy.

17
00:01:01,714 --> 00:01:04,959
I'm curious how many of you
have heard of SUDEP.

18
00:01:05,603 --> 00:01:09,969
This is a very well-educated audience,
and I see only a few hands.

19
00:01:09,993 --> 00:01:13,959
SUDEP is when an otherwise
healthy person with epilepsy

20
00:01:13,983 --> 00:01:17,873
dies and they can't attribute it
to anything they can find in an autopsy.

21
00:01:20,082 --> 00:01:23,685
There is a SUDEP
every seven to nine minutes.

22
00:01:24,217 --> 00:01:27,096
That's on average two per TED Talk.

23
00:01:31,411 --> 00:01:35,070
Now, a normal brain
has electrical activity.

24
00:01:35,094 --> 00:01:37,229
You can see some of the electrical waves

25
00:01:37,253 --> 00:01:40,317
coming out of this picture
of a brain here.

26
00:01:40,341 --> 00:01:44,087
And these should look
like typical electrical activity

27
00:01:44,111 --> 00:01:46,246
that an EEG could read on the surface.

28
00:01:46,270 --> 00:01:50,687
When you have a seizure,
it's a bit of unusual electrical activity,

29
00:01:50,711 --> 00:01:52,096
and it can be focal.

30
00:01:52,120 --> 00:01:54,620
It can take place
in just a small part of your brain.

31
00:01:54,644 --> 00:01:58,452
When that happens,
you might have a strange sensation.

32
00:01:58,476 --> 00:02:01,509
Several could be happening
here in the audience right now,

33
00:02:01,533 --> 00:02:04,135
and the person next to you
might not even know.

34
00:02:04,159 --> 00:02:08,079
However, if you have a seizure
where that little brush fire spreads

35
00:02:08,103 --> 00:02:10,095
like a forest fire over the brain,

36
00:02:10,119 --> 00:02:11,793
then it generalizes,

37
00:02:11,817 --> 00:02:16,388
and that generalized seizure
takes your consciousness away

38
00:02:16,412 --> 00:02:17,988
and causes you to convulse.

39
00:02:18,680 --> 00:02:22,822
There are more SUDEPs
in the United States every year

40
00:02:22,846 --> 00:02:25,505
than sudden infant death syndrome.

41
00:02:26,298 --> 00:02:29,300
Now, how many of you have heard
of sudden infant death syndrome?

42
00:02:29,324 --> 00:02:31,280
Right? Pretty much every hand goes up.

43
00:02:31,304 --> 00:02:33,027
So what's going on here?

44
00:02:33,051 --> 00:02:38,194
Why is this so much more common
and yet people haven't heard of it?

45
00:02:38,218 --> 00:02:40,654
And what can you do to prevent it?

46
00:02:40,678 --> 00:02:42,980
Well, there are two things,
scientifically shown,

47
00:02:43,004 --> 00:02:46,377
that prevent or reduce the risk of SUDEP.

48
00:02:47,115 --> 00:02:49,670
The first is: "Follow
your doctor's instructions,

49
00:02:49,694 --> 00:02:51,130
take your medications."

50
00:02:51,154 --> 00:02:53,355
Two-thirds of people who have epilepsy

51
00:02:53,379 --> 00:02:55,498
get it under control
with their medications.

52
00:02:55,848 --> 00:03:00,019
The second thing that reduces
the risk of SUDEP is companionship.

53
00:03:00,043 --> 00:03:04,737
It's having somebody there
at the time that you have a seizure.

54
00:03:04,761 --> 00:03:08,682
Now, SUDEP, even though
most of you have never heard of it,

55
00:03:08,706 --> 00:03:13,895
is actually the number two cause
of years of potential life lost

56
00:03:13,919 --> 00:03:16,895
of all neurological disorders.

57
00:03:17,197 --> 00:03:21,483
The vertical axis is the number of deaths

58
00:03:21,507 --> 00:03:24,673
times the remaining life span,

59
00:03:24,697 --> 00:03:28,136
so higher is much worse impact.

60
00:03:28,522 --> 00:03:31,496
SUDEP, however, unlike these others,

61
00:03:31,520 --> 00:03:37,236
is something that people right here
could do something to push that down.

62
00:03:37,778 --> 00:03:44,778
Now, what is Roz Picard, an AI researcher,
doing here telling you about SUDEP, right?

63
00:03:44,802 --> 00:03:47,323
I'm not a neurologist.

64
00:03:47,347 --> 00:03:51,562
When I was working at the Media Lab
on measurement of emotion,

65
00:03:51,586 --> 00:03:54,808
trying to make our machines
more intelligent about our emotions,

66
00:03:54,832 --> 00:03:57,363
we started doing a lot of work
measuring stress.

67
00:03:59,212 --> 00:04:01,148
We built lots of sensors

68
00:04:01,172 --> 00:04:03,553
that measured it
in lots of different ways.

69
00:04:03,577 --> 00:04:05,897
But one of them in particular

70
00:04:05,921 --> 00:04:10,033
grew out of some of this very old work
with measuring sweaty palms

71
00:04:10,057 --> 00:04:11,801
with an electrical signal.

72
00:04:11,825 --> 00:04:13,574
This is a signal of skin conductance

73
00:04:13,598 --> 00:04:15,716
that's known to go up
when you get nervous,

74
00:04:15,740 --> 00:04:19,312
but it turns out it also goes up with
a lot of other interesting conditions.

75
00:04:19,336 --> 00:04:22,358
But measuring it with wires on your hand
is really inconvenient.

76
00:04:22,382 --> 00:04:26,127
So we invented a bunch of other ways
of doing this at the MIT Media Lab.

77
00:04:26,151 --> 00:04:28,324
And with these wearables,

78
00:04:28,348 --> 00:04:33,153
we started to collect the first-ever
clinical quality data 24-7.

79
00:04:33,609 --> 00:04:36,142
Here's a picture of what that looked like

80
00:04:36,166 --> 00:04:42,761
the first time an MIT student collected
skin conductance on the wrist 24-7.

81
00:04:43,187 --> 00:04:45,545
Let's zoom in a little bit here.

82
00:04:45,569 --> 00:04:48,594
What you see is 24 hours
from left to right,

83
00:04:48,618 --> 00:04:50,545
and here is two days of data.

84
00:04:50,569 --> 00:04:53,251
And first, what surprised us

85
00:04:53,275 --> 00:04:56,621
was sleep was the biggest
peak of the day.

86
00:04:56,645 --> 00:04:58,395
Now, that sounds broken, right?

87
00:04:58,419 --> 00:05:02,597
You're calm when you're asleep,
so what's going on here?

88
00:05:02,621 --> 00:05:05,074
Well, it turns out
that our physiology during sleep

89
00:05:05,098 --> 00:05:07,741
is very different
than our physiology during wake,

90
00:05:07,765 --> 00:05:09,949
and while there's still a bit of a mystery

91
00:05:09,973 --> 00:05:13,717
why these peaks are usually
the biggest of the day during sleep,

92
00:05:13,741 --> 00:05:16,911
we now believe they're related
to memory consolidation

93
00:05:16,935 --> 00:05:18,990
and memory formation during sleep.

94
00:05:19,895 --> 00:05:23,466
We also saw things
that were exactly what we expected.

95
00:05:23,490 --> 00:05:25,840
When an MIT student
is working hard in the lab

96
00:05:25,864 --> 00:05:27,253
or on homeworks,

97
00:05:27,277 --> 00:05:31,824
there is not only emotional stress,
but there's cognitive load,

98
00:05:31,848 --> 00:05:36,709
and it turns out that cognitive load,
cognitive effort, mental engagement,

99
00:05:36,733 --> 00:05:38,927
excitement about learning something --

100
00:05:38,951 --> 00:05:41,288
those things also make the signal go up.

101
00:05:43,014 --> 00:05:46,752
Unfortunately, to the embarrassment
of we MIT professors,

102
00:05:46,776 --> 00:05:47,855
(Laughter)

103
00:05:47,879 --> 00:05:52,347
the low point every day
is classroom activity.

104
00:05:52,810 --> 00:05:55,326
Now, I am just showing you
one person's data here,

105
00:05:55,350 --> 00:05:57,960
but this, unfortunately,
is true in general.

106
00:06:00,400 --> 00:06:04,789
This sweatband has inside it
a homebuilt skin-conductance sensor,

107
00:06:04,813 --> 00:06:09,899
and one day, one of our undergrads
knocked on my door

108
00:06:09,923 --> 00:06:12,598
right at the end of the December semester,

109
00:06:12,622 --> 00:06:14,812
and he said, "Professor Picard,

110
00:06:14,836 --> 00:06:17,550
can I please borrow
one of your wristband sensors?

111
00:06:17,574 --> 00:06:21,303
My little brother has autism,
he can't talk,

112
00:06:21,327 --> 00:06:23,494
and I want to see
what's stressing him out."

113
00:06:24,434 --> 00:06:27,292
And I said, "Sure, in fact,
don't just take one, take two,"

114
00:06:27,316 --> 00:06:30,108
because they broke easily back then.

115
00:06:30,132 --> 00:06:32,752
So he took them home,
he put them on his little brother.

116
00:06:32,776 --> 00:06:35,945
Now, I was back in MIT,
looking at the data on my laptop,

117
00:06:35,969 --> 00:06:38,830
and the first day, I thought,
"Hmm, that's odd,

118
00:06:38,854 --> 00:06:41,845
he put them on both wrists
instead of waiting for one to break.

119
00:06:41,869 --> 00:06:44,189
OK, fine, don't follow my instructions."

120
00:06:44,213 --> 00:06:45,931
I'm glad he didn't.

121
00:06:46,377 --> 00:06:49,583
Second day -- chill.
Looked like classroom activity.

122
00:06:49,607 --> 00:06:50,869
(Laughter)

123
00:06:51,389 --> 00:06:52,718
A few more days ahead.

124
00:06:53,404 --> 00:06:57,918
The next day, one wrist signal was flat

125
00:06:57,942 --> 00:07:01,900
and the other had
the biggest peak I've ever seen,

126
00:07:02,496 --> 00:07:04,814
and I thought, "What's going on?

127
00:07:04,838 --> 00:07:08,298
We've stressed people out at MIT
every way imaginable.

128
00:07:08,933 --> 00:07:10,671
I've never seen a peak this big."

129
00:07:12,409 --> 00:07:14,378
And it was only on one side.

130
00:07:14,402 --> 00:07:17,546
How can you be stressed on one side
of your body and not the other?

131
00:07:17,570 --> 00:07:20,440
So I thought one or both sensors
must be broken.

132
00:07:21,185 --> 00:07:23,102
Now, I'm an electroengineer by training,

133
00:07:23,126 --> 00:07:25,795
so I started a whole bunch of stuff
to try to debug this,

134
00:07:25,819 --> 00:07:28,318
and long story short,
I could not reproduce this.

135
00:07:28,962 --> 00:07:31,995
So I resorted to old-fashioned debugging.

136
00:07:32,019 --> 00:07:35,301
I called the student at home on vacation.

137
00:07:35,325 --> 00:07:39,668
"Hi, how's your little brother?
How's your Christmas?

138
00:07:39,692 --> 00:07:42,973
Hey, do you have any idea
what happened to him?"

139
00:07:42,997 --> 00:07:44,953
And I gave this particular date and time,

140
00:07:44,977 --> 00:07:46,278
and the data.

141
00:07:46,302 --> 00:07:49,717
And he said, "I don't know,
I'll check the diary."

142
00:07:50,850 --> 00:07:53,758
Diary? An MIT student keeps a diary?

143
00:07:53,782 --> 00:07:55,601
So I waited and he came back.

144
00:07:55,625 --> 00:07:57,164
He had the exact date and time,

145
00:07:57,188 --> 00:08:01,041
and he says, "That was right before
he had a grand mal seizure."

146
00:08:02,942 --> 00:08:06,291
Now, at the time, I didn't know
anything about epilepsy,

147
00:08:06,315 --> 00:08:08,474
and did a bunch of research,

148
00:08:08,498 --> 00:08:11,624
realized that another student's dad
is chief of neurosurgery

149
00:08:11,648 --> 00:08:13,648
at Children's Hospital Boston,

150
00:08:13,672 --> 00:08:15,961
screwed up my courage
and called Dr. Joe Madsen.

151
00:08:15,985 --> 00:08:18,454
"Hi, Dr. Madsen,
my name's Rosalind Picard.

152
00:08:18,478 --> 00:08:22,374
Is it possible somebody could have

153
00:08:22,398 --> 00:08:26,801
a huge sympathetic
nervous system surge" --

154
00:08:26,825 --> 00:08:29,403
that's what drives the skin conductance --

155
00:08:29,427 --> 00:08:31,099
"20 minutes before a seizure?"

156
00:08:32,172 --> 00:08:34,291
And he says, "Probably not."

157
00:08:35,997 --> 00:08:37,331
He says, "It's interesting.

158
00:08:37,355 --> 00:08:40,123
We've had people whose hair
stands on end on one arm

159
00:08:40,147 --> 00:08:42,227
20 minutes before a seizure."

160
00:08:43,060 --> 00:08:44,375
And I'm like, "On one arm?"

161
00:08:44,399 --> 00:08:46,529
I didn't want to tell him that, initially,

162
00:08:46,553 --> 00:08:48,942
because I thought this was too ridiculous.

163
00:08:48,966 --> 00:08:51,259
He explained how this could
happen in the brain,

164
00:08:51,283 --> 00:08:53,456
and he got interested.
I showed him the data.

165
00:08:53,480 --> 00:08:56,428
We made a whole bunch more devices,
got them safety certified.

166
00:08:56,452 --> 00:08:58,526
90 families were being
enrolled in a study,

167
00:08:58,550 --> 00:09:01,779
all with children who were going
to be monitored 24-7

168
00:09:01,803 --> 00:09:05,174
with gold-standard EEG on their scalp

169
00:09:05,198 --> 00:09:06,741
for reading the brain activity,

170
00:09:06,765 --> 00:09:08,708
video to watch the behavior,

171
00:09:08,732 --> 00:09:12,392
electrocardiogram -- ECG --
and now EDA, electrodermal activity,

172
00:09:12,416 --> 00:09:15,001
to see if there was
something in this periphery

173
00:09:15,025 --> 00:09:17,471
that we could easily pick up,
related to a seizure.

174
00:09:18,263 --> 00:09:24,704
We found, in 100 percent
of the first batch of grand mal seizures,

175
00:09:24,728 --> 00:09:28,361
this whopper of responses
in the skin conductance.

176
00:09:28,385 --> 00:09:30,388
The blue in the middle, the boy's sleep,

177
00:09:30,412 --> 00:09:32,384
is usually the biggest peak of the day.

178
00:09:32,408 --> 00:09:36,040
These three seizures you see here
are popping out of the forest

179
00:09:36,064 --> 00:09:37,540
like redwood trees.

180
00:09:39,096 --> 00:09:42,676
Furthermore, when you couple
the skin conductance at the top

181
00:09:42,700 --> 00:09:46,217
with the movement from the wrist

182
00:09:46,241 --> 00:09:51,041
and you get lots of data
and train machine learning and AI on it,

183
00:09:51,065 --> 00:09:55,858
you can build an automated AI
that detects these patterns

184
00:09:55,882 --> 00:10:00,097
much better than just
a shake detector can do.

185
00:10:00,121 --> 00:10:04,388
So we realized that we needed
to get this out,

186
00:10:04,412 --> 00:10:06,770
and with the PhD work of Ming-Zher Poh

187
00:10:06,794 --> 00:10:09,821
and later great improvements by Empatica,

188
00:10:09,845 --> 00:10:13,689
this has made progress and the seizure
detection is much more accurate.

189
00:10:13,713 --> 00:10:16,618
But we also learned some other things
about SUDEP during this.

190
00:10:16,642 --> 00:10:20,054
One thing we learned is that SUDEP,

191
00:10:20,078 --> 00:10:22,990
while it's rare after
a generalized tonic-clonic seizure,

192
00:10:23,014 --> 00:10:26,251
that's when it's most likely
to happen -- after that type.

193
00:10:26,275 --> 00:10:29,394
And when it happens,
it doesn't happen during the seizure,

194
00:10:29,418 --> 00:10:32,284
and it doesn't usually happen
immediately afterwards,

195
00:10:32,308 --> 00:10:34,023
but immediately afterwards,

196
00:10:34,047 --> 00:10:37,109
when the person just seems
very still and quiet,

197
00:10:37,133 --> 00:10:42,080
they may go into another phase,
where the breathing stops,

198
00:10:42,104 --> 00:10:45,128
and then after the breathing stops,
later the heart stops.

199
00:10:45,152 --> 00:10:47,195
So there's some time
to get somebody there.

200
00:10:48,349 --> 00:10:53,253
We also learned that there is a region
deep in the brain called the amygdala,

201
00:10:53,277 --> 00:10:56,213
which we had been studying
in our emotion research a lot.

202
00:10:56,237 --> 00:10:57,518
We have two amygdalas,

203
00:10:57,542 --> 00:10:59,215
and if you stimulate the right one,

204
00:10:59,239 --> 00:11:01,978
you get a big right
skin conductance response.

205
00:11:02,002 --> 00:11:06,247
Now, you have to sign up right now
for a craniotomy to get this done,

206
00:11:06,271 --> 00:11:09,104
not exactly something
we're going to volunteer to do,

207
00:11:09,128 --> 00:11:11,641
but it causes a big right skin
conductance response.

208
00:11:11,665 --> 00:11:15,518
Stimulate the left one, big left
skin conductance response on the palm.

209
00:11:15,542 --> 00:11:19,582
And furthermore, when somebody
stimulates your amygdala

210
00:11:19,606 --> 00:11:23,025
while you're sitting there
and you might just be working,

211
00:11:23,049 --> 00:11:24,937
you don't show any signs of distress,

212
00:11:25,620 --> 00:11:26,906
but you stop breathing,

213
00:11:28,470 --> 00:11:31,787
and you don't start again
until somebody stimulates you.

214
00:11:32,599 --> 00:11:34,112
"Hey, Roz, are you there?"

215
00:11:34,136 --> 00:11:36,067
And you open your mouth to talk.

216
00:11:36,868 --> 00:11:39,062
As you take that breath to speak,

217
00:11:39,086 --> 00:11:40,591
you start breathing again.

218
00:11:42,762 --> 00:11:46,382
So we had started with work on stress,

219
00:11:46,406 --> 00:11:48,659
which had enabled us
to build lots of sensors

220
00:11:48,683 --> 00:11:50,815
that were gathering
high quality enough data

221
00:11:50,839 --> 00:11:53,760
that we could leave the lab
and start to get this in the wild;

222
00:11:53,784 --> 00:11:56,648
accidentally found a whopper
of a response with the seizure,

223
00:11:56,672 --> 00:11:59,531
neurological activation that can cause
a much bigger response

224
00:11:59,555 --> 00:12:00,865
than traditional stressors;

225
00:12:00,889 --> 00:12:04,087
lots of partnership with hospitals
and an epilepsy monitoring unit,

226
00:12:04,111 --> 00:12:05,939
especially Children's Hospital Boston

227
00:12:05,963 --> 00:12:07,114
and the Brigham;

228
00:12:07,138 --> 00:12:09,663
and machine learning and AI on top of this

229
00:12:09,687 --> 00:12:12,708
to take and collect lots more data

230
00:12:12,732 --> 00:12:15,361
in service of trying
to understand these events

231
00:12:15,385 --> 00:12:16,979
and if we could prevent SUDEP.

232
00:12:18,023 --> 00:12:21,586
This is now commercialized by Empatica,

233
00:12:21,610 --> 00:12:24,499
a start-up that I had
the privilege to cofound,

234
00:12:24,523 --> 00:12:28,729
and the team there has done an amazing job
improving the technology

235
00:12:28,753 --> 00:12:30,512
to make a very beautiful sensor

236
00:12:30,536 --> 00:12:34,395
that not only tells time and does steps
and sleep and all that good stuff,

237
00:12:34,419 --> 00:12:37,515
but this is running real-time
AI and machine learning

238
00:12:37,539 --> 00:12:40,157
to detect generalized
tonic-clonic seizures

239
00:12:40,181 --> 00:12:42,332
and send an alert for help

240
00:12:42,356 --> 00:12:45,548
if I were to have a seizure
and lose consciousness.

241
00:12:45,572 --> 00:12:48,388
This just got FDA-approved

242
00:12:48,412 --> 00:12:52,552
as the first smartwatch
to get approved in neurology.

243
00:12:53,923 --> 00:13:00,923
(Applause)

244
00:13:02,674 --> 00:13:06,205
Now, the next slide is what made
my skin conductance go up.

245
00:13:06,811 --> 00:13:08,517
One morning, I'm checking my email

246
00:13:08,541 --> 00:13:10,838
and I see a story from a mom

247
00:13:10,862 --> 00:13:12,748
who said she was in the shower,

248
00:13:12,772 --> 00:13:15,225
and her phone was
on the counter by the shower,

249
00:13:15,249 --> 00:13:17,406
and it said her daughter
might need her help.

250
00:13:18,196 --> 00:13:21,577
So she interrupts her shower and goes
running to her daughter's bedroom,

251
00:13:21,601 --> 00:13:25,344
and she finds her daughter
facedown in bed, blue and not breathing.

252
00:13:25,368 --> 00:13:28,649
She flips her over -- human stimulation --

253
00:13:28,673 --> 00:13:32,356
and her daughter takes a breath,
and another breath,

254
00:13:32,380 --> 00:13:36,334
and her daughter turns pink and is fine.

255
00:13:37,551 --> 00:13:40,775
I think I turned white reading this email.

256
00:13:40,799 --> 00:13:43,121
My first response is,
"Oh no, it's not perfect.

257
00:13:43,145 --> 00:13:45,470
The Bluetooth could break,
the battery could die.

258
00:13:45,494 --> 00:13:47,971
All these things could go wrong.
Don't rely on this."

259
00:13:47,995 --> 00:13:51,403
And she said, "It's OK.
I know no technology is perfect.

260
00:13:51,427 --> 00:13:53,590
None of us can always
be there all the time.

261
00:13:54,844 --> 00:13:59,186
But this, this device plus AI

262
00:13:59,210 --> 00:14:02,314
enabled me to get there in time
to save my daughter's life."

263
00:14:05,933 --> 00:14:07,839
Now, I've been mentioning children,

264
00:14:07,863 --> 00:14:13,570
but SUDEP peaks, actually,
among people in their 20s, 30s and 40s,

265
00:14:13,594 --> 00:14:15,410
and the next line I'm going to put up

266
00:14:15,434 --> 00:14:17,887
is probably going to make
some people uncomfortable,

267
00:14:17,911 --> 00:14:20,285
but it's less uncomfortable
than we'll all be

268
00:14:20,309 --> 00:14:23,379
if this list is extended
to somebody you know.

269
00:14:24,437 --> 00:14:26,775
Could this happen to somebody you know?

270
00:14:27,197 --> 00:14:29,939
And the reason I bring up
this uncomfortable question

271
00:14:29,963 --> 00:14:34,875
is because one in 26 of you
will have epilepsy at some point,

272
00:14:34,899 --> 00:14:37,134
and from what I've been learning,

273
00:14:37,158 --> 00:14:40,492
people with epilepsy often don't tell
their friends and their neighbors

274
00:14:40,516 --> 00:14:41,674
that they have it.

275
00:14:41,698 --> 00:14:46,619
So if you're willing to let them
use an AI or whatever

276
00:14:46,643 --> 00:14:51,057
to summon you in a moment
of possible need,

277
00:14:51,081 --> 00:14:52,699
if you would let them know that,

278
00:14:52,723 --> 00:14:55,077
you could make a difference in their life.

279
00:14:55,728 --> 00:14:58,579
Why do all this hard work to build AIs?

280
00:15:00,101 --> 00:15:01,302
A couple of reasons here:

281
00:15:01,326 --> 00:15:03,489
one is Natasha, the girl who lived,

282
00:15:03,513 --> 00:15:05,905
and her family wanted me
to tell you her name.

283
00:15:06,977 --> 00:15:08,589
Another is her family

284
00:15:08,613 --> 00:15:10,804
and the wonderful people out there

285
00:15:10,828 --> 00:15:13,653
who want to be there to support people
who have conditions

286
00:15:13,677 --> 00:15:16,970
that they've felt uncomfortable
in the past mentioning to others.

287
00:15:17,901 --> 00:15:19,774
And the other reason is all of you,

288
00:15:19,798 --> 00:15:25,130
because we have the opportunity
to shape the future of AI.

289
00:15:25,154 --> 00:15:27,511
We can actually change it,

290
00:15:27,535 --> 00:15:29,939
because we are the ones building it.

291
00:15:29,963 --> 00:15:31,940
So let's build AI

292
00:15:31,964 --> 00:15:35,256
that makes everybody's lives better.

293
00:15:35,982 --> 00:15:37,133
Thank you.

294
00:15:37,157 --> 00:15:41,571
(Applause)