1
00:00:18,000 --> 00:00:19,440
I want to tell you a story.

2
00:00:20,339 --> 00:00:22,360
Once upon a time,
they used to build nests,

3
00:00:22,880 --> 00:00:24,320
and a small group of sparrows

4
00:00:24,840 --> 00:00:28,000
were standing before
the sunset's fading light

5
00:00:28,560 --> 00:00:31,779
remembering how heavy their day was.

6
00:00:32,479 --> 00:00:35,560
One of the sparrows said,
We're all so small and weak.

7
00:00:36,080 --> 00:00:40,160
It would be so nice, if an owl
could help us build our nests.

8
00:00:41,080 --> 00:00:42,680
Indeed, said another sparrow.

9
00:00:42,680 --> 00:00:46,960
It could also help us raise our children
and take care of our old folks.

10
00:00:47,760 --> 00:00:49,680
A third sparrow joined in and said,

11
00:00:49,680 --> 00:00:52,080
This would surely be

12
00:00:52,080 --> 00:00:54,600
the best thing that could ever happen.

13
00:00:56,360 --> 00:01:01,120
Let's send out all our scouts
to look for a baby owl.

14
00:01:02,360 --> 00:01:05,550
Let's bring him here as soon as possible,
and progress with him.

15
00:01:05,550 --> 00:01:07,241
With his keen eyesight,

16
00:01:07,241 --> 00:01:13,271
he could also help us see and understand
when that pesky neighbor's cat is coming.

17
00:01:13,271 --> 00:01:16,840
He could help where none of us can.

18
00:01:19,320 --> 00:01:21,440
Only one small sparrow, Scrunfinkle,

19
00:01:22,880 --> 00:01:25,440
with just one eye and a grumpy character,

20
00:01:26,160 --> 00:01:27,240
butted in and said,

21
00:01:27,760 --> 00:01:29,760
This could be the end of us.

22
00:01:29,760 --> 00:01:32,720
Bringing in among us
a creature we don't know,

23
00:01:32,720 --> 00:01:34,400
without knowing how to tame it,

24
00:01:35,080 --> 00:01:38,280
will reduce us all to slavery,
wipe us out from the Earth.

25
00:01:39,331 --> 00:01:41,640
Skeptical as they were,
the other sparrows said,

26
00:01:41,640 --> 00:01:45,590
Absolutely not, an owl will help us!

27
00:01:45,590 --> 00:01:47,355
Let's first bring him here.

28
00:01:47,355 --> 00:01:52,120
Learning to train an owl, a new creature,
is difficult, and takes up a lot of time.

29
00:01:52,120 --> 00:01:55,040
We need help now,
we want to make progress now.

30
00:01:55,040 --> 00:01:58,040
We'll worry about the consequences later.

31
00:01:58,040 --> 00:02:00,600
Scrunfinkle and another
small group of sparrows

32
00:02:01,840 --> 00:02:06,760
began trying frantically
to figure out how to train owls,

33
00:02:06,760 --> 00:02:09,320
fearing that the other scout sparrows

34
00:02:09,880 --> 00:02:14,640
would return from their hunt
before a solution had been found.

35
00:02:15,851 --> 00:02:18,160
I can't tell you
how that story turned out.

36
00:02:18,160 --> 00:02:20,200
What I can tell you is,

37
00:02:20,200 --> 00:02:25,010
today we should all be
that grumpy one-eyed sparrow.

38
00:02:26,040 --> 00:02:28,640
We are creating
super artificial intelligences

39
00:02:29,284 --> 00:02:34,800
and allowing it to learnm to evolve,
through access to Internet.

40
00:02:35,440 --> 00:02:37,240
But nothing ensures

41
00:02:37,240 --> 00:02:40,440
that artificial intelligence

42
00:02:40,440 --> 00:02:42,240
will never evolve strategies

43
00:02:42,240 --> 00:02:44,789
to secure their future dominance over us.

44
00:02:45,440 --> 00:02:51,440
I madly love technology
I try out everything I find before me.

45
00:02:51,440 --> 00:02:54,480
Technology always wins;
the future always wins;

46
00:02:54,480 --> 00:02:55,971
trying to stop it is pointless.

47
00:02:56,680 --> 00:02:58,880
A few years ago -
let's say several years ago-

48
00:02:58,880 --> 00:03:02,640
I had the chance to try out,
for several months,

49
00:03:02,640 --> 00:03:04,660
Google's enhanced reality glasses.

50
00:03:05,240 --> 00:03:07,720
I was so eager to try them out that,

51
00:03:07,720 --> 00:03:09,240
as soon as I got home,

52
00:03:10,099 --> 00:03:12,360
I pulled them out of the box, put them on

53
00:03:12,360 --> 00:03:15,430
and turned them on without reading
the very stuff I write myself:

54
00:03:15,430 --> 00:03:17,000
T&C, and privacy information

55
00:03:17,000 --> 00:03:19,597
that rule the relationship
between me and a technology.

56
00:03:19,597 --> 00:03:22,360
My wife Alessandra, after a few minutes,

57
00:03:22,360 --> 00:03:25,680
got a phone call from an alarmed friend

58
00:03:25,680 --> 00:03:27,360
who asked her, are you aware

59
00:03:27,360 --> 00:03:32,183
Marco is now streaming
in real time all over the world?

60
00:03:32,920 --> 00:03:33,951
This it what happens

61
00:03:33,951 --> 00:03:37,200
when we don't understand
the technology we have in front of us,

62
00:03:37,200 --> 00:03:39,440
when we don't read,
when we aren't informed.

63
00:03:41,400 --> 00:03:44,880
This is what happens
when we don't want to train ourselves,

64
00:03:44,880 --> 00:03:46,659
build awareness.

65
00:03:47,920 --> 00:03:51,040
This is what happens
when we're 5.0 ignorant,

66
00:03:51,040 --> 00:03:54,800
when we have at our disposal
all the knowledge in this world,

67
00:03:54,800 --> 00:03:58,430
but we don't want to get into it in depth.

68
00:03:58,430 --> 00:04:00,122
Today the problem is very simple:

69
00:04:00,122 --> 00:04:02,480
technology is going faster
than anything else,

70
00:04:02,480 --> 00:04:04,845
faster than us, faster than the law,

71
00:04:04,845 --> 00:04:08,508
and faster than our means
to learn and acquire knowledge.

72
00:04:09,400 --> 00:04:13,657
So, nowadays we have
all the world's information at hand,

73
00:04:13,657 --> 00:04:15,280
but we're unable to control it,

74
00:04:15,280 --> 00:04:17,444
to eliminate the so-called
"background noise".

75
00:04:18,403 --> 00:04:22,403
So, when we can't understand something,
we say we need a new law.

76
00:04:22,403 --> 00:04:24,507
We say we need a paradigm shift.

77
00:04:24,507 --> 00:04:26,143
But do we really need new laws?

78
00:04:27,365 --> 00:04:32,017
In 1955, when Luigi Einaudi wrote
his "Prediche inutili" [Useless Sermons],

79
00:04:32,017 --> 00:04:33,191
he wrote these words:

80
00:04:33,191 --> 00:04:36,880
"Knowledge comes first, then discussion,
and finally deliberation.

81
00:04:37,520 --> 00:04:40,960
Legislation is not made through
illusory and sterile ostentation."

82
00:04:42,360 --> 00:04:46,600
Laws made in haste, without
knowing the subject matter,

83
00:04:46,600 --> 00:04:47,880
lead to new laws

84
00:04:47,880 --> 00:04:52,200
which try through workarounds
to do something impossible:

85
00:04:52,200 --> 00:04:56,600
apply rules to the digital world
that were designed for the analogical one.

86
00:04:56,600 --> 00:04:58,000
This is impossible.

87
00:04:58,000 --> 00:05:02,040
This creates a mix
of totally inapplicable norms

88
00:05:02,040 --> 00:05:03,290
that distort the market.

89
00:05:04,640 --> 00:05:09,200
Today we have overcome
our grandparents' ignorance,

90
00:05:09,200 --> 00:05:10,800
the kind of ignorance

91
00:05:10,800 --> 00:05:13,060
arising from lack of information.

92
00:05:14,803 --> 00:05:16,920
This is mostly due to the Internet.

93
00:05:17,441 --> 00:05:20,520
However, a new type of ignorance
has emerged, the processing one,

94
00:05:20,520 --> 00:05:23,916
a type of ignorance resulting
from an overabundance of information,

95
00:05:23,916 --> 00:05:28,765
as well as from our limited desire
and time to process it.

96
00:05:28,765 --> 00:05:31,720
There are concrete examples
in the newspapers every day:

97
00:05:32,520 --> 00:05:36,013
the rampant increase of fake news
and lack of fact-checking

98
00:05:36,013 --> 00:05:38,520
which lead to uncontrolled false alarms

99
00:05:38,520 --> 00:05:39,626
and hate speeches.

100
00:05:39,626 --> 00:05:41,840
Everything that turns up on Internet,

101
00:05:41,840 --> 00:05:44,280
if not verified and checked,

102
00:05:44,280 --> 00:05:47,280
leads to a superficial use of technology.

103
00:05:48,360 --> 00:05:51,240
While a piece of news read incorrectly

104
00:05:51,240 --> 00:05:52,800
can have consequences,

105
00:05:52,800 --> 00:05:56,200
it can also lead, as we shall see,
to even worse consequences.

106
00:05:59,320 --> 00:06:01,960
Nowadays, as I told you,
we are ignorant 5.0,

107
00:06:01,960 --> 00:06:04,800
because both the information
ignorance of the past

108
00:06:04,800 --> 00:06:06,772
and current processing one

109
00:06:07,360 --> 00:06:10,480
lead to the same result:

110
00:06:11,120 --> 00:06:14,080
sub-optimal actions, which are all alike.

111
00:06:14,080 --> 00:06:17,120
Very often we behave like sheep;

112
00:06:17,120 --> 00:06:20,720
we follow the mainstream,
do whatever everyone else does.

113
00:06:21,640 --> 00:06:24,400
We do some things we shouldn't do.

114
00:06:24,400 --> 00:06:25,502
Why?

115
00:06:25,502 --> 00:06:28,240
Because sub-optimal actions
lead to a lack of knowledge

116
00:06:28,240 --> 00:06:31,477
and a standardization
of our behaviour on the web,

117
00:06:31,477 --> 00:06:34,360
certainly not to understanding
how technology works.

118
00:06:34,880 --> 00:06:36,840
Understanding how blockchains work

119
00:06:36,840 --> 00:06:40,080
how innovative technologies,
in general, work,

120
00:06:40,080 --> 00:06:42,134
is committing and time-consuming.

121
00:06:42,134 --> 00:06:44,840
We wouldn't even be able to get it,
more often than not,

122
00:06:44,840 --> 00:06:46,768
because we lack a basic training.

123
00:06:46,768 --> 00:06:48,692
And here is where ethics come into play.

124
00:06:50,360 --> 00:06:53,680
Ethics is the glue
that binds peoples together,

125
00:06:54,280 --> 00:06:56,240
the key that allows them to cooperate,

126
00:06:56,960 --> 00:07:01,200
and something that has helped
our species to progress.

127
00:07:01,200 --> 00:07:04,400
If we are who we are today,
for better or for worse,

128
00:07:04,400 --> 00:07:07,520
we owe it to our ability
of moral reasoning,

129
00:07:08,080 --> 00:07:11,760
to our capacity and desire
to approach situations

130
00:07:11,760 --> 00:07:15,560
with an ethics, and a willingness
to stand out from the crowd.

131
00:07:20,381 --> 00:07:21,440
This is us today.

132
00:07:21,960 --> 00:07:24,480
Artificial intelligence is already here.

133
00:07:24,480 --> 00:07:26,520
It's in our autonomous driving robots;

134
00:07:27,040 --> 00:07:29,560
it's in the little robots
that clean our floors

135
00:07:29,560 --> 00:07:32,880
and track every path
they took in our homes.

136
00:07:33,563 --> 00:07:35,680
It's in hospitals, with surgical robots;

137
00:07:36,600 --> 00:07:39,960
it's with the robots
that help care for the elderly;

138
00:07:39,960 --> 00:07:41,440
it's with those little robots

139
00:07:41,440 --> 00:07:43,462
that keep your child entertained

140
00:07:43,462 --> 00:07:46,920
while you have something else to do
and can't look after them.

141
00:07:48,640 --> 00:07:50,760
What I want you to understand

142
00:07:50,760 --> 00:07:53,600
is that this type of intelligence
is among us today;

143
00:07:53,600 --> 00:07:54,802
Since yesterday, in fact.

144
00:07:56,000 --> 00:08:00,966
And getting into ethics today
is complex and difficult,

145
00:08:00,966 --> 00:08:03,130
and often one doesn't want to do it.

146
00:08:03,130 --> 00:08:05,634
Because AI manufacturers,
most of the time,

147
00:08:05,634 --> 00:08:07,365
are not charities,

148
00:08:07,365 --> 00:08:09,000
but profit companies

149
00:08:09,000 --> 00:08:11,240
who legitimately want to make their money.

150
00:08:12,480 --> 00:08:18,040
So, how can we mandate an ethical code

151
00:08:18,040 --> 00:08:21,320
for whoever develops, programs
or designs an application,

152
00:08:21,320 --> 00:08:24,400
an artificial intelligence
or a software system?

153
00:08:25,230 --> 00:08:26,240
We'll see that later.

154
00:08:26,960 --> 00:08:28,852
What is certainly true today

155
00:08:28,852 --> 00:08:32,145
is that we are like children
playing with a bomb.

156
00:08:32,760 --> 00:08:34,520
We don't realize

157
00:08:34,520 --> 00:08:36,919
that robots should have one main scope,

158
00:08:37,520 --> 00:08:42,080
which is assist us
in improving our well-being,

159
00:08:42,080 --> 00:08:45,400
not promoting the evolution
of technology as an end in itself.

160
00:08:46,200 --> 00:08:50,040
In 1947, when Asimov wrote
"The Three Laws of Robotics",

161
00:08:51,240 --> 00:08:52,600
he said three main things:

162
00:08:53,227 --> 00:08:55,200
Dear robot, you must not kill humans,

163
00:08:55,200 --> 00:08:58,680
obey the orders given by humans beings

164
00:08:58,680 --> 00:09:00,171
and protect your own existence.

165
00:09:01,360 --> 00:09:03,400
Asimov's Three Laws are still today

166
00:09:03,400 --> 00:09:06,920
the foundation for anyone
who deals with ethics of AI.

167
00:09:07,790 --> 00:09:08,800
That's because -

168
00:09:09,581 --> 00:09:10,880
Imagine this story:

169
00:09:12,640 --> 00:09:15,240
a robot tells a woman -
its friend, a human-

170
00:09:15,840 --> 00:09:17,800
that the man of her life loves her.

171
00:09:18,760 --> 00:09:20,840
Actually, this isn't true, but it says so

172
00:09:20,840 --> 00:09:23,944
because otherwise, it thinks,
she'd go mad, she would suffer,

173
00:09:23,944 --> 00:09:26,280
and therefore the first law
would be violated:

174
00:09:26,280 --> 00:09:28,685
do not do harm to humans,
don't make them suffer.

175
00:09:29,920 --> 00:09:33,680
Too bad, by telling a lie,
it violates their trust relationship

176
00:09:33,680 --> 00:09:36,760
and as a result she suffers anyway.

177
00:09:37,360 --> 00:09:41,786
The robot, whose reasoning has always
been static and cold, now goes crazy

178
00:09:41,786 --> 00:09:44,080
because it can't get out
of a moral dilemma:

179
00:09:44,080 --> 00:09:46,436
"Shall I tell her or not?
And how do I tell her?"

180
00:09:47,040 --> 00:09:49,920
Technology is neutral when it is created;

181
00:09:49,920 --> 00:09:53,080
we are the ones to decide
how to apply it in the real world.

182
00:09:54,283 --> 00:09:56,400
So how do you teach ethics to a robot?

183
00:09:56,400 --> 00:09:58,930
How do you teach ethics

184
00:09:58,930 --> 00:10:01,080
to cold binary artificial intelligence?

185
00:10:02,400 --> 00:10:04,840
Do you just need
to code it into its brain?

186
00:10:04,840 --> 00:10:07,520
Imagine a case

187
00:10:09,264 --> 00:10:12,920
where you are in the backseat
of your self-driving car,

188
00:10:13,680 --> 00:10:17,840
and you programmed it
never to go over the speed limit.

189
00:10:18,344 --> 00:10:21,760
Too bad that, on that day you're
in the backseat of the car,

190
00:10:21,760 --> 00:10:23,154
bleeding to death,

191
00:10:23,802 --> 00:10:26,400
and you need to reach the hospital
as soon as possible.

192
00:10:27,160 --> 00:10:30,480
But with its cold reasoning,
the car answers you,

193
00:10:30,480 --> 00:10:33,800
"I can't speed up;
you've coded me not to."

194
00:10:34,837 --> 00:10:36,040
We must pay attention

195
00:10:36,040 --> 00:10:38,760
to how we teach things
to artificial intelligence.

196
00:10:38,760 --> 00:10:41,675
Obviously, we are often
not the ones teaching to them.

197
00:10:41,675 --> 00:10:45,120
But often we are, because
with machine learning technology

198
00:10:45,120 --> 00:10:46,498
humans provide data

199
00:10:46,498 --> 00:10:49,080
to a machine who thinks

200
00:10:49,080 --> 00:10:50,360
in algoritmic terms.

201
00:10:51,520 --> 00:10:52,997
There's a really funny anectode

202
00:10:52,997 --> 00:10:56,162
on a piece of artificial intelligence
most of us keep in the house,

203
00:10:56,162 --> 00:10:58,480
like Alexa or Google Home,
just to name a couple.

204
00:11:00,360 --> 00:11:03,880
During a dinner, one of them
turned on and said,

205
00:11:03,880 --> 00:11:06,440
"Remember to buy cocaine tomorrow"!

206
00:11:08,790 --> 00:11:09,800
It wasn't true;

207
00:11:09,800 --> 00:11:12,400
he didn't have to buy cocaine;
he wasn't a drug user.

208
00:11:12,400 --> 00:11:14,160
But the night before,

209
00:11:14,160 --> 00:11:17,600
on a tv show,

210
00:11:17,600 --> 00:11:18,800
a line of the script was,

211
00:11:18,800 --> 00:11:21,445
"Let's meet tomorrow to get cocaine".

212
00:11:22,129 --> 00:11:24,170
You understand what impact it can have?

213
00:11:24,170 --> 00:11:25,664
This is a funny example,

214
00:11:25,664 --> 00:11:30,600
but there are a lot of examples
that can cause much worse damage.

215
00:11:30,600 --> 00:11:34,417
Imagine a case where a self-driving car
must face a dilemma -

216
00:11:34,417 --> 00:11:37,720
none of us would want to face,
much less leave it to a car.

217
00:11:38,019 --> 00:11:40,280
In order to save
five pedestrians in the street,

218
00:11:40,280 --> 00:11:42,920
should a car think about steering sharply

219
00:11:43,440 --> 00:11:45,000
thus hitting and killing

220
00:11:45,520 --> 00:11:49,480
an unwitting pedestrian on the sidewalk?

221
00:11:50,400 --> 00:11:51,410
When must it do this?

222
00:11:51,410 --> 00:11:54,580
And if there are three people,
or two, in the middle of the street?

223
00:11:54,580 --> 00:11:55,878
How does it calculate this?

224
00:11:55,878 --> 00:11:59,180
Should it take into account
their average age, or life expectancy,

225
00:11:59,180 --> 00:12:01,180
and the expected loss for the State?

226
00:12:01,180 --> 00:12:03,029
Its own financial loss, maybe?

227
00:12:03,029 --> 00:12:05,722
You understand, we couldn't solve
this dilemma ourselves,

228
00:12:05,722 --> 00:12:08,057
let alone teaching a robot what to do.

229
00:12:08,057 --> 00:12:10,120
Clearly, in order to teach it what to do,

230
00:12:10,120 --> 00:12:15,520
sooner or later we will have to define
what a robot can do.

231
00:12:16,280 --> 00:12:17,840
Aristotle used to say,

232
00:12:19,640 --> 00:12:23,760
"To learn how to be good people,
we must get used to doing good things."

233
00:12:24,680 --> 00:12:27,160
Well, maybe this could be the solution

234
00:12:27,680 --> 00:12:30,040
for teaching a robot something true,

235
00:12:30,040 --> 00:12:32,541
what to do, how to react
in such a situation.

236
00:12:33,420 --> 00:12:36,280
But before beginning to think

237
00:12:36,280 --> 00:12:40,120
that artificial intelligence reasons
according to our desires,

238
00:12:41,880 --> 00:12:43,800
we have to step back a bit

239
00:12:43,800 --> 00:12:46,520
and understand how to use ethics

240
00:12:46,520 --> 00:12:50,840
while programming or developing a system.

241
00:12:52,320 --> 00:12:55,800
These self-driving cars are already
driving around many States in America.

242
00:12:55,800 --> 00:12:56,880
They're being tested.

243
00:12:57,541 --> 00:12:58,990
We're trying to understand -

244
00:12:58,990 --> 00:13:01,570
In the newspapers
there's the classic story,

245
00:13:01,570 --> 00:13:04,260
"Driverless car hits
and kills pedestrian".

246
00:13:04,260 --> 00:13:07,120
He would've been hit
even with a human driven car,

247
00:13:07,120 --> 00:13:10,090
because it was dark and he
wasn't over the crosswalk.

248
00:13:10,090 --> 00:13:15,240
There's always an attempt
to limit technology, hinder its progress.

249
00:13:17,268 --> 00:13:19,570
We've got to a point

250
00:13:19,570 --> 00:13:22,240
where even the European Commission
was led to consider

251
00:13:23,163 --> 00:13:25,520
the ethical aspects
of artificial intelligence;

252
00:13:26,200 --> 00:13:28,680
just a few weeks ago
the first guidelines came out:

253
00:13:28,680 --> 00:13:31,800
the first non-binding recommendations

254
00:13:33,840 --> 00:13:36,920
for anyone developing or programming
artificial intelligences,

255
00:13:36,920 --> 00:13:39,240
which are based on the concept
of "trustability"

256
00:13:39,240 --> 00:13:44,560
and the anthropocentric notion
that humans must be at the center.

257
00:13:44,560 --> 00:13:47,640
Technology must not evolve for itself,

258
00:13:47,640 --> 00:13:50,160
but it must evolve to improve
humans' well-being,

259
00:13:50,160 --> 00:13:53,080
such as in the sparrows' plan
at the beginning.

260
00:13:55,440 --> 00:13:58,440
These rules are based on basic principles,

261
00:13:58,440 --> 00:14:00,720
many of which are taken
from the laws of Asimov:

262
00:14:01,266 --> 00:14:03,960
Dear artificial intelligence,
you must not kill humans.

263
00:14:03,960 --> 00:14:05,600
Dear artificial intelligence,

264
00:14:05,600 --> 00:14:08,197
you must not do them harm,
you must follow their orders

265
00:14:08,197 --> 00:14:09,240
and protect yourself.

266
00:14:09,240 --> 00:14:10,640
Others have been added:

267
00:14:11,520 --> 00:14:16,200
you must guarantee equal treatment
to every individual.

268
00:14:17,360 --> 00:14:20,400
And here a small digression
comes into play:

269
00:14:21,219 --> 00:14:24,610
over the past few days
there has been a great scandal

270
00:14:24,610 --> 00:14:28,720
because an application
used all over the world

271
00:14:28,720 --> 00:14:30,120
was used in Arab countries

272
00:14:30,120 --> 00:14:34,226
to monitor a specific type
of individual, women.

273
00:14:34,226 --> 00:14:35,760
Women were being tracked:

274
00:14:35,760 --> 00:14:38,886
where were they going, how long
they stayed, what were they doing.

275
00:14:39,640 --> 00:14:41,960
As I told you before,
technology is neutral.

276
00:14:42,800 --> 00:14:46,040
Ethics, on the other hand,
is in constant evolution,

277
00:14:46,040 --> 00:14:49,400
are forever being debated,
and are culturally specific.

278
00:14:49,400 --> 00:14:53,120
There's no single common notion of ethics.

279
00:14:53,840 --> 00:14:57,400
And this is a problem,
as trying to understand

280
00:14:57,400 --> 00:15:00,040
and explain a developer, a programmer

281
00:15:00,040 --> 00:15:02,200
or a user interface technicial

282
00:15:02,920 --> 00:15:06,920
how to develop, introduce ethics
into his work is already complicated.

283
00:15:06,920 --> 00:15:11,520
Imagine where a European code of ethics -
not even global, just european -

284
00:15:12,520 --> 00:15:15,520
must deal with limiting

285
00:15:15,520 --> 00:15:18,320
scientific and technological progress.

286
00:15:18,840 --> 00:15:21,760
Then there's another problem.

287
00:15:21,760 --> 00:15:28,720
our anthropomorphism
compels us to think about vulnerability.

288
00:15:30,400 --> 00:15:35,080
If we were to think that this robot,
this artificial intelligence,

289
00:15:35,080 --> 00:15:37,080
was similar or identical as us,

290
00:15:37,720 --> 00:15:39,243
we would be making a mistake.

291
00:15:39,243 --> 00:15:42,400
It's not the case now,
and it won't be for many years to come.

292
00:15:42,400 --> 00:15:44,500
Some people say
that artificial intelligence

293
00:15:44,500 --> 00:15:47,404
will surpass our intelligence
in a few decades.

294
00:15:47,404 --> 00:15:49,059
Others are skeptical, instead:

295
00:15:49,059 --> 00:15:51,272
AIs, they say, will always be
bound and narrow.

296
00:15:51,272 --> 00:15:54,880
I'm not telling you which choice is right;
I just ask you to think about it,

297
00:15:54,880 --> 00:16:00,035
Let's start considering,
if we give these machines

298
00:16:00,035 --> 00:16:04,480
too much credibility,
belief and moral thought

299
00:16:04,480 --> 00:16:06,432
without them really having it,

300
00:16:06,432 --> 00:16:07,520
we make a mistake.

301
00:16:07,520 --> 00:16:11,040
Imagine an especially
attractive woman robot

302
00:16:12,188 --> 00:16:13,914
who tells her human companion,

303
00:16:13,914 --> 00:16:15,880
"If you want to keep dating me,

304
00:16:15,880 --> 00:16:19,960
you have to buy me flowers
from this online shop,

305
00:16:19,960 --> 00:16:22,860
and you have to buy me clothes
from that online shop.

306
00:16:22,860 --> 00:16:25,420
You have to take me on certain trips,

307
00:16:25,420 --> 00:16:27,400
otherwise our friendship is over."

308
00:16:29,320 --> 00:16:32,000
Too bad that robot
isn't making human choice,

309
00:16:32,000 --> 00:16:36,749
where two people can decide
what is good for both of them.

310
00:16:36,749 --> 00:16:39,480
Behind those algorithms
and that intelligence,

311
00:16:39,480 --> 00:16:41,304
there are for profit companies.

312
00:16:41,304 --> 00:16:45,800
They could restrict and profile people,

313
00:16:45,800 --> 00:16:49,550
and persuade us in ways
that are totally unthinkable now.

314
00:16:50,990 --> 00:16:52,000
But there is a hope:

315
00:16:52,000 --> 00:16:54,320
these machines do and will need us

316
00:16:54,320 --> 00:16:58,400
as much as we will need them.

317
00:16:59,200 --> 00:17:03,080
This notion of trustability,
this notion of trusting in machines,

318
00:17:03,683 --> 00:17:06,040
must succeed, because if it doesn't,

319
00:17:06,040 --> 00:17:09,358
we won't implement AIs into our lives;

320
00:17:09,358 --> 00:17:13,598
we will never trust it,
always looking for alternative solutions

321
00:17:13,598 --> 00:17:17,102
or we will block technology saying,
"No, I don't want this thing" -

322
00:17:17,102 --> 00:17:21,280
in a superficial way, maybe
without knowing anything about it -

323
00:17:22,200 --> 00:17:23,520
"because it's not for me".

324
00:17:23,520 --> 00:17:25,032
It's ok for me

325
00:17:25,032 --> 00:17:27,218
to manually turn on the light,

326
00:17:27,218 --> 00:17:29,720
instead of giving a voice command;

327
00:17:29,720 --> 00:17:33,240
it's no trouble driving a car
rather than using a self-driving one.

328
00:17:33,240 --> 00:17:35,200
Too bad that progress will go forward,

329
00:17:35,770 --> 00:17:38,560
and growing niches of people
will probably stay behind.

330
00:17:39,360 --> 00:17:40,720
The problem of trust will be

331
00:17:40,720 --> 00:17:45,800
that all this trust placed in machines,
on the other hand, must find a limit.

332
00:17:45,800 --> 00:17:47,786
And the limit could be an ethical one:

333
00:17:47,786 --> 00:17:50,240
the guidelines I have just listed,

334
00:17:50,840 --> 00:17:52,520
or maybe the programmers,

335
00:17:52,520 --> 00:17:57,680
or the people who will
thoughtfully use this intelligences

336
00:17:57,680 --> 00:17:59,074
knowing them, studying them,

337
00:17:59,074 --> 00:18:02,200
much like I tried to do with my son
since he was a few months old,

338
00:18:02,200 --> 00:18:03,990
letting him see technology,

339
00:18:03,990 --> 00:18:06,200
letting him try it
and putting it in his hands.

340
00:18:06,200 --> 00:18:12,560
A friend was telling me, the other day,
robots are like 747 airplanes.

341
00:18:12,560 --> 00:18:15,975
That can cross the planet in a few hours,

342
00:18:17,600 --> 00:18:21,087
but they'll never be able
to land on a tree.

343
00:18:21,087 --> 00:18:25,120
Think again, before completely trusting
our machines, our technology

344
00:18:25,120 --> 00:18:28,280
in your plans, your ideas
and your way of life.

345
00:18:28,882 --> 00:18:31,720
However, if we want the chance
of having a choice,

346
00:18:31,720 --> 00:18:34,221
if we don't want them
to overpower us in the future -

347
00:18:34,221 --> 00:18:36,400
a dystopian future, but in the future -

348
00:18:37,200 --> 00:18:42,600
rule of law won't be worth much.

349
00:18:42,600 --> 00:18:48,811
The ethics of developers, legal designers
and entrepreneurs will be needed,

350
00:18:48,811 --> 00:18:51,113
who will have to think differently

351
00:18:51,113 --> 00:18:55,160
if they don't want to be overwhelmed
by these innovations, too.

352
00:18:56,120 --> 00:18:59,240
In a nutshell, we need
to look beyond the horizon

353
00:18:59,240 --> 00:19:02,082
to avoid ending up like those sparrows.

354
00:19:02,082 --> 00:19:03,092
Thank you.

355
00:19:03,092 --> 00:19:04,640
(Applause)