What will a future without secrets look like? | Alessandro Acquisti

202,219 views ・ 2013-10-18

TED


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

00:12
I would like to tell you a story
0
12641
2354
00:14
connecting the notorious privacy incident
1
14995
3176
00:18
involving Adam and Eve,
2
18171
2769
00:20
and the remarkable shift in the boundaries
3
20940
3446
00:24
between public and private which has occurred
4
24386
2686
00:27
in the past 10 years.
5
27072
1770
00:28
You know the incident.
6
28842
1298
00:30
Adam and Eve one day in the Garden of Eden
7
30140
3330
00:33
realize they are naked.
8
33470
1843
00:35
They freak out.
9
35313
1500
00:36
And the rest is history.
10
36813
2757
00:39
Nowadays, Adam and Eve
11
39570
2188
00:41
would probably act differently.
12
41758
2361
00:44
[@Adam Last nite was a blast! loved dat apple LOL]
13
44119
2268
00:46
[@Eve yep.. babe, know what happened to my pants tho?]
14
46387
1873
00:48
We do reveal so much more information
15
48260
2636
00:50
about ourselves online than ever before,
16
50896
3334
00:54
and so much information about us
17
54230
1704
00:55
is being collected by organizations.
18
55934
2224
00:58
Now there is much to gain and benefit
19
58158
3282
01:01
from this massive analysis of personal information,
20
61440
2446
01:03
or big data,
21
63886
1946
01:05
but there are also complex tradeoffs that come
22
65832
2638
01:08
from giving away our privacy.
23
68470
3098
01:11
And my story is about these tradeoffs.
24
71568
4023
01:15
We start with an observation which, in my mind,
25
75591
2584
01:18
has become clearer and clearer in the past few years,
26
78175
3327
01:21
that any personal information
27
81502
2097
01:23
can become sensitive information.
28
83599
2285
01:25
Back in the year 2000, about 100 billion photos
29
85884
4125
01:30
were shot worldwide,
30
90009
1912
01:31
but only a minuscule proportion of them
31
91921
3065
01:34
were actually uploaded online.
32
94986
1883
01:36
In 2010, only on Facebook, in a single month,
33
96869
3361
01:40
2.5 billion photos were uploaded,
34
100230
3270
01:43
most of them identified.
35
103500
1882
01:45
In the same span of time,
36
105382
1880
01:47
computers' ability to recognize people in photos
37
107262
4870
01:52
improved by three orders of magnitude.
38
112132
3608
01:55
What happens when you combine
39
115740
1882
01:57
these technologies together:
40
117622
1501
01:59
increasing availability of facial data;
41
119123
2658
02:01
improving facial recognizing ability by computers;
42
121781
3648
02:05
but also cloud computing,
43
125429
2182
02:07
which gives anyone in this theater
44
127611
1888
02:09
the kind of computational power
45
129499
1560
02:11
which a few years ago was only the domain
46
131059
1886
02:12
of three-letter agencies;
47
132945
1782
02:14
and ubiquitous computing,
48
134727
1378
02:16
which allows my phone, which is not a supercomputer,
49
136105
2892
02:18
to connect to the Internet
50
138997
1671
02:20
and do there hundreds of thousands
51
140668
2334
02:23
of face metrics in a few seconds?
52
143002
2639
02:25
Well, we conjecture that the result
53
145641
2628
02:28
of this combination of technologies
54
148269
2064
02:30
will be a radical change in our very notions
55
150333
2888
02:33
of privacy and anonymity.
56
153221
2257
02:35
To test that, we did an experiment
57
155478
1993
02:37
on Carnegie Mellon University campus.
58
157471
2121
02:39
We asked students who were walking by
59
159592
2099
02:41
to participate in a study,
60
161691
1779
02:43
and we took a shot with a webcam,
61
163470
2562
02:46
and we asked them to fill out a survey on a laptop.
62
166032
2782
02:48
While they were filling out the survey,
63
168814
1979
02:50
we uploaded their shot to a cloud-computing cluster,
64
170793
2797
02:53
and we started using a facial recognizer
65
173590
1727
02:55
to match that shot to a database
66
175317
2405
02:57
of some hundreds of thousands of images
67
177722
2393
03:00
which we had downloaded from Facebook profiles.
68
180115
3596
03:03
By the time the subject reached the last page
69
183711
3259
03:06
on the survey, the page had been dynamically updated
70
186970
3347
03:10
with the 10 best matching photos
71
190317
2313
03:12
which the recognizer had found,
72
192630
2285
03:14
and we asked the subjects to indicate
73
194915
1738
03:16
whether he or she found themselves in the photo.
74
196653
4120
03:20
Do you see the subject?
75
200773
3699
03:24
Well, the computer did, and in fact did so
76
204472
2845
03:27
for one out of three subjects.
77
207317
2149
03:29
So essentially, we can start from an anonymous face,
78
209466
3184
03:32
offline or online, and we can use facial recognition
79
212650
3484
03:36
to give a name to that anonymous face
80
216134
2360
03:38
thanks to social media data.
81
218494
2108
03:40
But a few years back, we did something else.
82
220602
1872
03:42
We started from social media data,
83
222474
1823
03:44
we combined it statistically with data
84
224297
3051
03:47
from U.S. government social security,
85
227348
2102
03:49
and we ended up predicting social security numbers,
86
229450
3324
03:52
which in the United States
87
232774
1512
03:54
are extremely sensitive information.
88
234286
2040
03:56
Do you see where I'm going with this?
89
236326
2093
03:58
So if you combine the two studies together,
90
238419
2922
04:01
then the question becomes,
91
241341
1512
04:02
can you start from a face and,
92
242853
2720
04:05
using facial recognition, find a name
93
245573
2311
04:07
and publicly available information
94
247884
2669
04:10
about that name and that person,
95
250553
1932
04:12
and from that publicly available information
96
252485
2248
04:14
infer non-publicly available information,
97
254733
2042
04:16
much more sensitive ones
98
256775
1606
04:18
which you link back to the face?
99
258381
1492
04:19
And the answer is, yes, we can, and we did.
100
259873
1916
04:21
Of course, the accuracy keeps getting worse.
101
261789
2568
04:24
[27% of subjects' first 5 SSN digits identified (with 4 attempts)]
102
264357
944
04:25
But in fact, we even decided to develop an iPhone app
103
265301
3827
04:29
which uses the phone's internal camera
104
269128
2715
04:31
to take a shot of a subject
105
271843
1600
04:33
and then upload it to a cloud
106
273443
1487
04:34
and then do what I just described to you in real time:
107
274930
2662
04:37
looking for a match, finding public information,
108
277592
2088
04:39
trying to infer sensitive information,
109
279680
1730
04:41
and then sending back to the phone
110
281410
2591
04:44
so that it is overlaid on the face of the subject,
111
284001
3609
04:47
an example of augmented reality,
112
287610
1901
04:49
probably a creepy example of augmented reality.
113
289511
2451
04:51
In fact, we didn't develop the app to make it available,
114
291962
3339
04:55
just as a proof of concept.
115
295301
1922
04:57
In fact, take these technologies
116
297223
2313
04:59
and push them to their logical extreme.
117
299536
1837
05:01
Imagine a future in which strangers around you
118
301373
2719
05:04
will look at you through their Google Glasses
119
304092
2311
05:06
or, one day, their contact lenses,
120
306403
2307
05:08
and use seven or eight data points about you
121
308710
4020
05:12
to infer anything else
122
312730
2582
05:15
which may be known about you.
123
315312
2603
05:17
What will this future without secrets look like?
124
317915
4794
05:22
And should we care?
125
322709
1964
05:24
We may like to believe
126
324673
1891
05:26
that the future with so much wealth of data
127
326564
3040
05:29
would be a future with no more biases,
128
329604
2514
05:32
but in fact, having so much information
129
332118
3583
05:35
doesn't mean that we will make decisions
130
335701
2191
05:37
which are more objective.
131
337892
1706
05:39
In another experiment, we presented to our subjects
132
339598
2560
05:42
information about a potential job candidate.
133
342158
2246
05:44
We included in this information some references
134
344404
3178
05:47
to some funny, absolutely legal,
135
347582
2646
05:50
but perhaps slightly embarrassing information
136
350228
2465
05:52
that the subject had posted online.
137
352693
2020
05:54
Now interestingly, among our subjects,
138
354713
2366
05:57
some had posted comparable information,
139
357079
3083
06:00
and some had not.
140
360162
2362
06:02
Which group do you think
141
362524
1949
06:04
was more likely to judge harshly our subject?
142
364473
4552
06:09
Paradoxically, it was the group
143
369025
1957
06:10
who had posted similar information,
144
370982
1733
06:12
an example of moral dissonance.
145
372715
2942
06:15
Now you may be thinking,
146
375657
1750
06:17
this does not apply to me,
147
377407
1702
06:19
because I have nothing to hide.
148
379109
2162
06:21
But in fact, privacy is not about
149
381271
2482
06:23
having something negative to hide.
150
383753
3676
06:27
Imagine that you are the H.R. director
151
387429
2354
06:29
of a certain organization, and you receive résumés,
152
389783
2947
06:32
and you decide to find more information about the candidates.
153
392730
2473
06:35
Therefore, you Google their names
154
395203
2460
06:37
and in a certain universe,
155
397663
2240
06:39
you find this information.
156
399903
2008
06:41
Or in a parallel universe, you find this information.
157
401911
4437
06:46
Do you think that you would be equally likely
158
406348
2717
06:49
to call either candidate for an interview?
159
409065
2803
06:51
If you think so, then you are not
160
411868
2282
06:54
like the U.S. employers who are, in fact,
161
414150
2582
06:56
part of our experiment, meaning we did exactly that.
162
416732
3307
07:00
We created Facebook profiles, manipulating traits,
163
420039
3182
07:03
then we started sending out résumés to companies in the U.S.,
164
423221
2851
07:06
and we detected, we monitored,
165
426072
1908
07:07
whether they were searching for our candidates,
166
427980
2393
07:10
and whether they were acting on the information
167
430373
1832
07:12
they found on social media. And they were.
168
432205
1938
07:14
Discrimination was happening through social media
169
434143
2101
07:16
for equally skilled candidates.
170
436244
3073
07:19
Now marketers like us to believe
171
439317
4575
07:23
that all information about us will always
172
443892
2269
07:26
be used in a manner which is in our favor.
173
446161
3273
07:29
But think again. Why should that be always the case?
174
449434
3715
07:33
In a movie which came out a few years ago,
175
453149
2664
07:35
"Minority Report," a famous scene
176
455813
2553
07:38
had Tom Cruise walk in a mall
177
458366
2576
07:40
and holographic personalized advertising
178
460942
3776
07:44
would appear around him.
179
464718
1835
07:46
Now, that movie is set in 2054,
180
466553
3227
07:49
about 40 years from now,
181
469780
1642
07:51
and as exciting as that technology looks,
182
471422
2908
07:54
it already vastly underestimates
183
474330
2646
07:56
the amount of information that organizations
184
476976
2140
07:59
can gather about you, and how they can use it
185
479116
2483
08:01
to influence you in a way that you will not even detect.
186
481599
3398
08:04
So as an example, this is another experiment
187
484997
2103
08:07
actually we are running, not yet completed.
188
487100
2273
08:09
Imagine that an organization has access
189
489373
2319
08:11
to your list of Facebook friends,
190
491692
2056
08:13
and through some kind of algorithm
191
493748
1772
08:15
they can detect the two friends that you like the most.
192
495520
3734
08:19
And then they create, in real time,
193
499254
2280
08:21
a facial composite of these two friends.
194
501534
2842
08:24
Now studies prior to ours have shown that people
195
504376
3069
08:27
don't recognize any longer even themselves
196
507445
2885
08:30
in facial composites, but they react
197
510330
2462
08:32
to those composites in a positive manner.
198
512792
2117
08:34
So next time you are looking for a certain product,
199
514909
3415
08:38
and there is an ad suggesting you to buy it,
200
518324
2559
08:40
it will not be just a standard spokesperson.
201
520883
2907
08:43
It will be one of your friends,
202
523790
2313
08:46
and you will not even know that this is happening.
203
526103
3303
08:49
Now the problem is that
204
529406
2413
08:51
the current policy mechanisms we have
205
531819
2519
08:54
to protect ourselves from the abuses of personal information
206
534338
3438
08:57
are like bringing a knife to a gunfight.
207
537776
2984
09:00
One of these mechanisms is transparency,
208
540760
2913
09:03
telling people what you are going to do with their data.
209
543673
3200
09:06
And in principle, that's a very good thing.
210
546873
2106
09:08
It's necessary, but it is not sufficient.
211
548979
3667
09:12
Transparency can be misdirected.
212
552646
3698
09:16
You can tell people what you are going to do,
213
556344
2104
09:18
and then you still nudge them to disclose
214
558448
2232
09:20
arbitrary amounts of personal information.
215
560680
2623
09:23
So in yet another experiment, this one with students,
216
563303
2886
09:26
we asked them to provide information
217
566189
3058
09:29
about their campus behavior,
218
569247
1813
09:31
including pretty sensitive questions, such as this one.
219
571060
2940
09:34
[Have you ever cheated in an exam?]
220
574000
621
09:34
Now to one group of subjects, we told them,
221
574621
2300
09:36
"Only other students will see your answers."
222
576921
2841
09:39
To another group of subjects, we told them,
223
579762
1579
09:41
"Students and faculty will see your answers."
224
581341
3561
09:44
Transparency. Notification. And sure enough, this worked,
225
584902
2591
09:47
in the sense that the first group of subjects
226
587493
1407
09:48
were much more likely to disclose than the second.
227
588900
2568
09:51
It makes sense, right?
228
591468
1520
09:52
But then we added the misdirection.
229
592988
1490
09:54
We repeated the experiment with the same two groups,
230
594478
2760
09:57
this time adding a delay
231
597238
2427
09:59
between the time we told subjects
232
599665
2935
10:02
how we would use their data
233
602600
2080
10:04
and the time we actually started answering the questions.
234
604680
4388
10:09
How long a delay do you think we had to add
235
609068
2561
10:11
in order to nullify the inhibitory effect
236
611629
4613
10:16
of knowing that faculty would see your answers?
237
616242
3411
10:19
Ten minutes?
238
619653
1780
10:21
Five minutes?
239
621433
1791
10:23
One minute?
240
623224
1776
10:25
How about 15 seconds?
241
625000
2049
10:27
Fifteen seconds were sufficient to have the two groups
242
627049
2668
10:29
disclose the same amount of information,
243
629717
1568
10:31
as if the second group now no longer cares
244
631285
2746
10:34
for faculty reading their answers.
245
634031
2656
10:36
Now I have to admit that this talk so far
246
636687
3336
10:40
may sound exceedingly gloomy,
247
640023
2480
10:42
but that is not my point.
248
642503
1721
10:44
In fact, I want to share with you the fact that
249
644224
2699
10:46
there are alternatives.
250
646923
1772
10:48
The way we are doing things now is not the only way
251
648695
2499
10:51
they can done, and certainly not the best way
252
651194
3037
10:54
they can be done.
253
654231
2027
10:56
When someone tells you, "People don't care about privacy,"
254
656258
4171
11:00
consider whether the game has been designed
255
660429
2642
11:03
and rigged so that they cannot care about privacy,
256
663071
2724
11:05
and coming to the realization that these manipulations occur
257
665795
3262
11:09
is already halfway through the process
258
669057
1607
11:10
of being able to protect yourself.
259
670664
2258
11:12
When someone tells you that privacy is incompatible
260
672922
3710
11:16
with the benefits of big data,
261
676632
1849
11:18
consider that in the last 20 years,
262
678481
2473
11:20
researchers have created technologies
263
680954
1917
11:22
to allow virtually any electronic transactions
264
682871
3318
11:26
to take place in a more privacy-preserving manner.
265
686189
3749
11:29
We can browse the Internet anonymously.
266
689938
2555
11:32
We can send emails that can only be read
267
692493
2678
11:35
by the intended recipient, not even the NSA.
268
695171
3709
11:38
We can have even privacy-preserving data mining.
269
698880
2997
11:41
In other words, we can have the benefits of big data
270
701877
3894
11:45
while protecting privacy.
271
705771
2132
11:47
Of course, these technologies imply a shifting
272
707903
3791
11:51
of cost and revenues
273
711694
1546
11:53
between data holders and data subjects,
274
713240
2107
11:55
which is why, perhaps, you don't hear more about them.
275
715347
3453
11:58
Which brings me back to the Garden of Eden.
276
718800
3706
12:02
There is a second privacy interpretation
277
722506
2780
12:05
of the story of the Garden of Eden
278
725286
1809
12:07
which doesn't have to do with the issue
279
727095
2096
12:09
of Adam and Eve feeling naked
280
729191
2225
12:11
and feeling ashamed.
281
731416
2381
12:13
You can find echoes of this interpretation
282
733797
2781
12:16
in John Milton's "Paradise Lost."
283
736578
2782
12:19
In the garden, Adam and Eve are materially content.
284
739360
4197
12:23
They're happy. They are satisfied.
285
743557
2104
12:25
However, they also lack knowledge
286
745661
2293
12:27
and self-awareness.
287
747954
1640
12:29
The moment they eat the aptly named
288
749594
3319
12:32
fruit of knowledge,
289
752913
1293
12:34
that's when they discover themselves.
290
754206
2605
12:36
They become aware. They achieve autonomy.
291
756811
4031
12:40
The price to pay, however, is leaving the garden.
292
760842
3126
12:43
So privacy, in a way, is both the means
293
763968
3881
12:47
and the price to pay for freedom.
294
767849
2962
12:50
Again, marketers tell us
295
770811
2770
12:53
that big data and social media
296
773581
3019
12:56
are not just a paradise of profit for them,
297
776600
2979
12:59
but a Garden of Eden for the rest of us.
298
779579
2457
13:02
We get free content.
299
782036
1238
13:03
We get to play Angry Birds. We get targeted apps.
300
783274
3123
13:06
But in fact, in a few years, organizations
301
786397
2897
13:09
will know so much about us,
302
789294
1609
13:10
they will be able to infer our desires
303
790903
2710
13:13
before we even form them, and perhaps
304
793613
2204
13:15
buy products on our behalf
305
795817
2447
13:18
before we even know we need them.
306
798264
2274
13:20
Now there was one English author
307
800538
3237
13:23
who anticipated this kind of future
308
803775
3045
13:26
where we would trade away
309
806820
1405
13:28
our autonomy and freedom for comfort.
310
808225
3548
13:31
Even more so than George Orwell,
311
811773
2161
13:33
the author is, of course, Aldous Huxley.
312
813934
2761
13:36
In "Brave New World," he imagines a society
313
816695
2854
13:39
where technologies that we created
314
819549
2171
13:41
originally for freedom
315
821720
1859
13:43
end up coercing us.
316
823579
2567
13:46
However, in the book, he also offers us a way out
317
826146
4791
13:50
of that society, similar to the path
318
830937
3438
13:54
that Adam and Eve had to follow to leave the garden.
319
834375
3955
13:58
In the words of the Savage,
320
838330
2147
14:00
regaining autonomy and freedom is possible,
321
840477
3069
14:03
although the price to pay is steep.
322
843546
2679
14:06
So I do believe that one of the defining fights
323
846225
5715
14:11
of our times will be the fight
324
851940
2563
14:14
for the control over personal information,
325
854503
2387
14:16
the fight over whether big data will become a force
326
856890
3507
14:20
for freedom,
327
860397
1289
14:21
rather than a force which will hiddenly manipulate us.
328
861686
4746
14:26
Right now, many of us
329
866432
2593
14:29
do not even know that the fight is going on,
330
869025
2753
14:31
but it is, whether you like it or not.
331
871778
2672
14:34
And at the risk of playing the serpent,
332
874450
2804
14:37
I will tell you that the tools for the fight
333
877254
2897
14:40
are here, the awareness of what is going on,
334
880151
3009
14:43
and in your hands,
335
883160
1355
14:44
just a few clicks away.
336
884515
3740
14:48
Thank you.
337
888255
1482
14:49
(Applause)
338
889737
4477
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