How AI can bring on a second Industrial Revolution | Kevin Kelly

340,981 views ・ 2017-01-12

TED


Please double-click on the English subtitles below to play the video.

00:00
Translator: Leslie Gauthier Reviewer: Camille Martínez
0
0
7000
00:14
I'm going to talk a little bit about where technology's going.
1
14966
3817
00:19
And often technology comes to us,
2
19509
2671
00:22
we're surprised by what it brings.
3
22566
1865
00:24
But there's actually a large aspect of technology
4
24455
3683
00:28
that's much more predictable,
5
28162
1802
00:29
and that's because technological systems of all sorts have leanings,
6
29988
4088
00:34
they have urgencies,
7
34100
1175
00:35
they have tendencies.
8
35299
1561
00:36
And those tendencies are derived from the very nature of the physics,
9
36884
4932
00:41
chemistry of wires and switches and electrons,
10
41840
3150
00:45
and they will make reoccurring patterns again and again.
11
45659
3602
00:49
And so those patterns produce these tendencies, these leanings.
12
49745
4874
00:54
You can almost think of it as sort of like gravity.
13
54643
2831
00:57
Imagine raindrops falling into a valley.
14
57498
2319
00:59
The actual path of a raindrop as it goes down the valley
15
59841
3088
01:02
is unpredictable.
16
62953
1169
01:04
We cannot see where it's going,
17
64651
1518
01:06
but the general direction is very inevitable:
18
66193
2277
01:08
it's downward.
19
68494
1234
01:10
And so these baked-in tendencies and urgencies
20
70377
4572
01:14
in technological systems
21
74973
1476
01:17
give us a sense of where things are going at the large form.
22
77051
3609
01:21
So in a large sense,
23
81149
1401
01:22
I would say that telephones were inevitable,
24
82574
3361
01:27
but the iPhone was not.
25
87005
1342
01:29
The Internet was inevitable,
26
89094
1478
01:31
but Twitter was not.
27
91274
1286
01:33
So we have many ongoing tendencies right now,
28
93036
3928
01:36
and I think one of the chief among them
29
96988
2720
01:39
is this tendency to make things smarter and smarter.
30
99732
3722
01:44
I call it cognifying -- cognification --
31
104041
2212
01:46
also known as artificial intelligence, or AI.
32
106783
2782
01:50
And I think that's going to be one of the most influential developments
33
110025
3746
01:53
and trends and directions and drives in our society in the next 20 years.
34
113795
5575
02:00
So, of course, it's already here.
35
120021
1985
02:02
We already have AI,
36
122030
2204
02:04
and often it works in the background,
37
124258
2398
02:06
in the back offices of hospitals,
38
126680
1586
02:08
where it's used to diagnose X-rays better than a human doctor.
39
128290
4686
02:13
It's in legal offices,
40
133000
1726
02:14
where it's used to go through legal evidence
41
134750
2368
02:17
better than a human paralawyer.
42
137142
1855
02:19
It's used to fly the plane that you came here with.
43
139506
3656
02:24
Human pilots only flew it seven to eight minutes,
44
144165
2381
02:26
the rest of the time the AI was driving.
45
146570
1953
02:28
And of course, in Netflix and Amazon,
46
148547
2173
02:30
it's in the background, making those recommendations.
47
150744
2530
02:33
That's what we have today.
48
153298
1261
02:34
And we have an example, of course, in a more front-facing aspect of it,
49
154583
4801
02:39
with the win of the AlphaGo, who beat the world's greatest Go champion.
50
159408
6629
02:46
But it's more than that.
51
166478
4053
02:50
If you play a video game, you're playing against an AI.
52
170555
2642
02:53
But recently, Google taught their AI
53
173221
4538
02:57
to actually learn how to play video games.
54
177783
2412
03:00
Again, teaching video games was already done,
55
180686
2709
03:03
but learning how to play a video game is another step.
56
183419
3897
03:07
That's artificial smartness.
57
187340
1678
03:10
What we're doing is taking this artificial smartness
58
190571
4522
03:15
and we're making it smarter and smarter.
59
195117
2423
03:18
There are three aspects to this general trend
60
198710
3895
03:22
that I think are underappreciated;
61
202629
1689
03:24
I think we would understand AI a lot better
62
204342
2277
03:26
if we understood these three things.
63
206643
2301
03:28
I think these things also would help us embrace AI,
64
208968
3283
03:32
because it's only by embracing it that we actually can steer it.
65
212275
3008
03:35
We can actually steer the specifics by embracing the larger trend.
66
215887
3157
03:39
So let me talk about those three different aspects.
67
219467
2979
03:42
The first one is: our own intelligence has a very poor understanding
68
222470
3673
03:46
of what intelligence is.
69
226167
1490
03:48
We tend to think of intelligence as a single dimension,
70
228110
3653
03:51
that it's kind of like a note that gets louder and louder.
71
231787
2750
03:54
It starts like with IQ measurement.
72
234561
2607
03:57
It starts with maybe a simple low IQ in a rat or mouse,
73
237192
4092
04:01
and maybe there's more in a chimpanzee,
74
241308
2134
04:03
and then maybe there's more in a stupid person,
75
243887
2191
04:06
and then maybe an average person like myself,
76
246102
2096
04:08
and then maybe a genius.
77
248222
1290
04:09
And this single IQ intelligence is getting greater and greater.
78
249536
4433
04:14
That's completely wrong.
79
254516
1151
04:15
That's not what intelligence is -- not what human intelligence is, anyway.
80
255691
3608
04:19
It's much more like a symphony of different notes,
81
259673
4506
04:24
and each of these notes is played on a different instrument of cognition.
82
264203
3609
04:27
There are many types of intelligences in our own minds.
83
267836
3701
04:31
We have deductive reasoning,
84
271561
3048
04:34
we have emotional intelligence,
85
274633
2221
04:36
we have spatial intelligence;
86
276878
1393
04:38
we have maybe 100 different types that are all grouped together,
87
278295
4021
04:42
and they vary in different strengths with different people.
88
282340
3905
04:46
And of course, if we go to animals, they also have another basket --
89
286269
4526
04:50
another symphony of different kinds of intelligences,
90
290819
2541
04:53
and sometimes those same instruments are the same that we have.
91
293384
3566
04:56
They can think in the same way, but they may have a different arrangement,
92
296974
3561
05:00
and maybe they're higher in some cases than humans,
93
300559
2467
05:03
like long-term memory in a squirrel is actually phenomenal,
94
303050
2837
05:05
so it can remember where it buried its nuts.
95
305911
2287
05:08
But in other cases they may be lower.
96
308222
1987
05:10
When we go to make machines,
97
310233
2730
05:12
we're going to engineer them in the same way,
98
312987
2196
05:15
where we'll make some of those types of smartness much greater than ours,
99
315207
5010
05:20
and many of them won't be anywhere near ours,
100
320241
2571
05:22
because they're not needed.
101
322836
1544
05:24
So we're going to take these things,
102
324404
2203
05:26
these artificial clusters,
103
326631
2081
05:28
and we'll be adding more varieties of artificial cognition to our AIs.
104
328736
5362
05:34
We're going to make them very, very specific.
105
334507
4071
05:38
So your calculator is smarter than you are in arithmetic already;
106
338602
6542
05:45
your GPS is smarter than you are in spatial navigation;
107
345168
3697
05:49
Google, Bing, are smarter than you are in long-term memory.
108
349337
4258
05:54
And we're going to take, again, these kinds of different types of thinking
109
354339
4530
05:58
and we'll put them into, like, a car.
110
358893
1933
06:00
The reason why we want to put them in a car so the car drives,
111
360850
3057
06:03
is because it's not driving like a human.
112
363931
2302
06:06
It's not thinking like us.
113
366257
1396
06:07
That's the whole feature of it.
114
367677
1920
06:09
It's not being distracted,
115
369621
1535
06:11
it's not worrying about whether it left the stove on,
116
371180
2754
06:13
or whether it should have majored in finance.
117
373958
2138
06:16
It's just driving.
118
376120
1153
06:17
(Laughter)
119
377297
1142
06:18
Just driving, OK?
120
378463
1841
06:20
And we actually might even come to advertise these
121
380328
2937
06:23
as "consciousness-free."
122
383289
1545
06:24
They're without consciousness,
123
384858
1774
06:26
they're not concerned about those things,
124
386656
2104
06:28
they're not distracted.
125
388784
1156
06:29
So in general, what we're trying to do
126
389964
2966
06:32
is make as many different types of thinking as we can.
127
392954
4500
06:37
We're going to populate the space
128
397804
2083
06:39
of all the different possible types, or species, of thinking.
129
399911
4159
06:44
And there actually may be some problems
130
404094
2068
06:46
that are so difficult in business and science
131
406186
2800
06:49
that our own type of human thinking may not be able to solve them alone.
132
409010
4042
06:53
We may need a two-step program,
133
413076
1992
06:55
which is to invent new kinds of thinking
134
415092
4203
06:59
that we can work alongside of to solve these really large problems,
135
419692
3734
07:03
say, like dark energy or quantum gravity.
136
423450
2918
07:08
What we're doing is making alien intelligences.
137
428496
2646
07:11
You might even think of this as, sort of, artificial aliens
138
431166
4069
07:15
in some senses.
139
435259
1207
07:16
And they're going to help us think different,
140
436490
2300
07:18
because thinking different is the engine of creation
141
438814
3632
07:22
and wealth and new economy.
142
442470
1867
07:25
The second aspect of this is that we are going to use AI
143
445835
4923
07:30
to basically make a second Industrial Revolution.
144
450782
2950
07:34
The first Industrial Revolution was based on the fact
145
454135
2773
07:36
that we invented something I would call artificial power.
146
456932
3462
07:40
Previous to that,
147
460879
1150
07:42
during the Agricultural Revolution,
148
462053
2034
07:44
everything that was made had to be made with human muscle
149
464111
3702
07:47
or animal power.
150
467837
1307
07:49
That was the only way to get anything done.
151
469565
2063
07:51
The great innovation during the Industrial Revolution was,
152
471652
2945
07:54
we harnessed steam power, fossil fuels,
153
474621
3109
07:57
to make this artificial power that we could use
154
477754
3856
08:01
to do anything we wanted to do.
155
481634
1669
08:03
So today when you drive down the highway,
156
483327
2772
08:06
you are, with a flick of the switch, commanding 250 horses --
157
486571
4525
08:11
250 horsepower --
158
491120
1572
08:12
which we can use to build skyscrapers, to build cities, to build roads,
159
492716
4692
08:17
to make factories that would churn out lines of chairs or refrigerators
160
497432
5789
08:23
way beyond our own power.
161
503245
1654
08:24
And that artificial power can also be distributed on wires on a grid
162
504923
6111
08:31
to every home, factory, farmstead,
163
511058
3199
08:34
and anybody could buy that artificial power,
164
514281
4191
08:38
just by plugging something in.
165
518496
1472
08:39
So this was a source of innovation as well,
166
519992
2439
08:42
because a farmer could take a manual hand pump,
167
522455
3418
08:45
and they could add this artificial power, this electricity,
168
525897
2916
08:48
and he'd have an electric pump.
169
528837
1497
08:50
And you multiply that by thousands or tens of thousands of times,
170
530358
3318
08:53
and that formula was what brought us the Industrial Revolution.
171
533700
3159
08:56
All the things that we see, all this progress that we now enjoy,
172
536883
3585
09:00
has come from the fact that we've done that.
173
540492
2063
09:02
We're going to do the same thing now with AI.
174
542579
2348
09:04
We're going to distribute that on a grid,
175
544951
2075
09:07
and now you can take that electric pump.
176
547050
2374
09:09
You can add some artificial intelligence,
177
549448
2968
09:12
and now you have a smart pump.
178
552440
1481
09:13
And that, multiplied by a million times,
179
553945
1928
09:15
is going to be this second Industrial Revolution.
180
555897
2363
09:18
So now the car is going down the highway,
181
558284
2382
09:20
it's 250 horsepower, but in addition, it's 250 minds.
182
560690
4294
09:25
That's the auto-driven car.
183
565008
1769
09:26
It's like a new commodity;
184
566801
1389
09:28
it's a new utility.
185
568214
1303
09:29
The AI is going to flow across the grid -- the cloud --
186
569541
3041
09:32
in the same way electricity did.
187
572606
1567
09:34
So everything that we had electrified,
188
574197
2380
09:36
we're now going to cognify.
189
576601
1723
09:38
And I would suggest, then,
190
578693
1385
09:40
that the formula for the next 10,000 start-ups
191
580102
3732
09:43
is very, very simple,
192
583858
1162
09:45
which is to take x and add AI.
193
585044
3167
09:49
That is the formula, that's what we're going to be doing.
194
589100
2812
09:51
And that is the way in which we're going to make
195
591936
3306
09:55
this second Industrial Revolution.
196
595266
1858
09:57
And by the way -- right now, this minute,
197
597148
2154
09:59
you can log on to Google
198
599326
1169
10:00
and you can purchase AI for six cents, 100 hits.
199
600519
3882
10:04
That's available right now.
200
604758
1604
10:06
So the third aspect of this
201
606386
2286
10:09
is that when we take this AI and embody it,
202
609315
2678
10:12
we get robots.
203
612017
1173
10:13
And robots are going to be bots,
204
613214
1703
10:14
they're going to be doing many of the tasks that we have already done.
205
614941
3328
10:20
A job is just a bunch of tasks,
206
620357
1528
10:21
so they're going to redefine our jobs
207
621909
1762
10:23
because they're going to do some of those tasks.
208
623695
2259
10:25
But they're also going to create whole new categories,
209
625978
3197
10:29
a whole new slew of tasks
210
629199
2247
10:31
that we didn't know we wanted to do before.
211
631470
2457
10:33
They're going to actually engender new kinds of jobs,
212
633951
3637
10:37
new kinds of tasks that we want done,
213
637612
2271
10:39
just as automation made up a whole bunch of new things
214
639907
3405
10:43
that we didn't know we needed before,
215
643336
1834
10:45
and now we can't live without them.
216
645194
1956
10:47
So they're going to produce even more jobs than they take away,
217
647174
3956
10:51
but it's important that a lot of the tasks that we're going to give them
218
651154
3434
10:54
are tasks that can be defined in terms of efficiency or productivity.
219
654612
4572
10:59
If you can specify a task,
220
659676
1828
11:01
either manual or conceptual,
221
661528
2235
11:03
that can be specified in terms of efficiency or productivity,
222
663787
4780
11:08
that goes to the bots.
223
668591
1777
11:10
Productivity is for robots.
224
670758
2178
11:12
What we're really good at is basically wasting time.
225
672960
3070
11:16
(Laughter)
226
676054
1028
11:17
We're really good at things that are inefficient.
227
677106
2316
11:19
Science is inherently inefficient.
228
679446
3025
11:22
It runs on that fact that you have one failure after another.
229
682816
2906
11:25
It runs on the fact that you make tests and experiments that don't work,
230
685746
3424
11:29
otherwise you're not learning.
231
689194
1442
11:30
It runs on the fact
232
690660
1162
11:31
that there is not a lot of efficiency in it.
233
691846
2083
11:33
Innovation by definition is inefficient,
234
693953
2779
11:36
because you make prototypes,
235
696756
1391
11:38
because you try stuff that fails, that doesn't work.
236
698171
2707
11:40
Exploration is inherently inefficiency.
237
700902
3112
11:44
Art is not efficient.
238
704038
1531
11:45
Human relationships are not efficient.
239
705593
2127
11:47
These are all the kinds of things we're going to gravitate to,
240
707744
2940
11:50
because they're not efficient.
241
710708
1475
11:52
Efficiency is for robots.
242
712207
2315
11:55
We're also going to learn that we're going to work with these AIs
243
715338
4123
11:59
because they think differently than us.
244
719485
1997
12:02
When Deep Blue beat the world's best chess champion,
245
722005
4314
12:06
people thought it was the end of chess.
246
726343
1929
12:08
But actually, it turns out that today, the best chess champion in the world
247
728296
4402
12:12
is not an AI.
248
732722
1557
12:14
And it's not a human.
249
734906
1181
12:16
It's the team of a human and an AI.
250
736111
2715
12:18
The best medical diagnostician is not a doctor, it's not an AI,
251
738850
4000
12:22
it's the team.
252
742874
1176
12:24
We're going to be working with these AIs,
253
744074
2149
12:26
and I think you'll be paid in the future
254
746247
1995
12:28
by how well you work with these bots.
255
748266
2391
12:31
So that's the third thing, is that they're different,
256
751026
4257
12:35
they're utility
257
755307
1165
12:36
and they are going to be something we work with rather than against.
258
756496
3816
12:40
We're working with these rather than against them.
259
760336
2639
12:42
So, the future:
260
762999
1477
12:44
Where does that take us?
261
764500
1420
12:45
I think that 25 years from now, they'll look back
262
765944
3567
12:49
and look at our understanding of AI and say,
263
769535
3125
12:52
"You didn't have AI. In fact, you didn't even have the Internet yet,
264
772684
3300
12:56
compared to what we're going to have 25 years from now."
265
776008
2741
12:59
There are no AI experts right now.
266
779849
3047
13:02
There's a lot of money going to it,
267
782920
1699
13:04
there are billions of dollars being spent on it;
268
784643
2268
13:06
it's a huge business,
269
786935
2164
13:09
but there are no experts, compared to what we'll know 20 years from now.
270
789123
4272
13:14
So we are just at the beginning of the beginning,
271
794064
2885
13:16
we're in the first hour of all this.
272
796973
2163
13:19
We're in the first hour of the Internet.
273
799160
1935
13:21
We're in the first hour of what's coming.
274
801119
2040
13:23
The most popular AI product in 20 years from now,
275
803183
4153
13:27
that everybody uses,
276
807360
1444
13:29
has not been invented yet.
277
809499
1544
13:32
That means that you're not late.
278
812464
2467
13:35
Thank you.
279
815684
1151
13:36
(Laughter)
280
816859
1026
13:37
(Applause)
281
817909
2757
About this website

This site will introduce you to YouTube videos that are useful for learning English. You will see English lessons taught by top-notch teachers from around the world. Double-click on the English subtitles displayed on each video page to play the video from there. The subtitles scroll in sync with the video playback. If you have any comments or requests, please contact us using this contact form.

https://forms.gle/WvT1wiN1qDtmnspy7