Jennifer Golbeck: The curly fry conundrum: Why social media "likes" say more than you might think

378,002 views ใƒป 2014-04-03

TED


ืื ื ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ืœืžื˜ื” ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ.

ืžืชืจื’ื: Ido Dekkers ืžื‘ืงืจ: Tal Dekkers
00:12
If you remember that first decade of the web,
0
12738
1997
ืื ืืชื ื–ื•ื›ืจื™ื ืืช ื”ืขืฉื•ืจ ื”ืจืืฉื•ืŸ ืฉืœ ื”ืื™ื ื˜ืจื ื˜,
00:14
it was really a static place.
1
14735
2255
ื–ื” ื”ื™ื” ื‘ืืžืช ืžืงื•ื ืกื˜ื˜ื™.
00:16
You could go online, you could look at pages,
2
16990
2245
ื™ื›ื•ืœืชื ืœื”ื™ื›ื ืก ืœืจืฉืช, ืœื”ืกืชื›ืœ ืขืœ ื“ืคื™ื,
00:19
and they were put up either by organizations
3
19235
2513
ื•ื”ื ื”ื•ืขืœื• ืื• ืขืœ ื™ื“ื™ ืืจื’ื•ื ื™ื
00:21
who had teams to do it
4
21748
1521
ืฉื”ื™ื• ืœื”ื ืฆื•ื•ืชื™ื ืฉืขืฉื• ืืช ื–ื”
00:23
or by individuals who were really tech-savvy
5
23269
2229
ืื• ืขืœ ื™ื“ื™ ืื ืฉื™ื ืคืจื˜ื™ื™ื ืฉื”ื™ื• ืžืžืฉ ื˜ื›ื ื•ืœื’ื™ื™ื
00:25
for the time.
6
25498
1737
ื‘ืื•ืชื• ื”ื–ืžืŸ.
00:27
And with the rise of social media
7
27235
1575
ื•ืขื ื”ืขืœื™ื” ืฉืœ ื”ืžื“ื™ื” ื”ื—ื‘ืจืชื™ืช
00:28
and social networks in the early 2000s,
8
28810
2399
ื•ื”ืจืฉืชื•ืช ื”ื—ื‘ืจืชื™ื•ืช ื‘ืชื—ื™ืœืช ืฉื ื•ืช ื” 2000,
00:31
the web was completely changed
9
31209
2149
ื”ืจืฉืช ื”ืฉืชื ืชื” ืœื’ืžืจื™
00:33
to a place where now the vast majority of content
10
33358
3608
ืœืžืงื•ื ืฉื‘ื• ืขื›ืฉื™ื• ืจื•ื‘ ื”ืชื•ื›ืŸ
00:36
we interact with is put up by average users,
11
36966
3312
ืื•ืชื• ืื ื• ืฆื•ืจื›ื™ื ืžื•ืขืœื” ืขืœ ื™ื“ื™ ื”ืžืฉืชืžืฉ ื”ืžืžื•ืฆืข,
00:40
either in YouTube videos or blog posts
12
40278
2697
ื‘ื™ืŸ ืื ื‘ืกืจื˜ื•ื ื™ ื™ื•ื˜ื™ื•ื‘ ืื• ืคื•ืกื˜ื™ื ื‘ื‘ืœื•ื’
00:42
or product reviews or social media postings.
13
42975
3315
ืื• ื‘ื™ืงื•ืจื•ืช ืžื•ืฆืจื™ื ืื• ืคื•ืกื˜ื™ื ื‘ืžื“ื™ื” ื”ื—ื‘ืจืชื™ืช.
00:46
And it's also become a much more interactive place,
14
46290
2347
ื•ื–ื” ื’ื ื”ืคืš ืœืžืงื•ื ื”ืจื‘ื” ื™ื•ืชืจ ืื™ื ื˜ืจืืงื˜ื™ื‘ื™,
00:48
where people are interacting with others,
15
48637
2637
ืฉื ืื ืฉื™ื ืžืชืงืฉืจื™ื ืขื ืื—ืจื™ื,
00:51
they're commenting, they're sharing,
16
51274
1696
ื”ื ืžื’ื™ื‘ื™ื, ื”ื ื—ื•ืœืงื™ื,
00:52
they're not just reading.
17
52970
1614
ื”ื ืœื ืจืง ืงื•ืจืื™ื.
00:54
So Facebook is not the only place you can do this,
18
54584
1866
ืื– ืคื™ื™ืกื‘ื•ืง ื”ื•ื ืœื ื”ืžืงื•ื ื”ื™ื—ื™ื“ ืœืขืฉื•ืช ื–ืืช.
00:56
but it's the biggest,
19
56450
1098
ืื‘ืœ ื”ื•ื ื”ื’ื“ื•ืœ ื‘ื™ื•ืชืจ,
00:57
and it serves to illustrate the numbers.
20
57548
1784
ื•ื”ื•ื ืžืฉืžืฉ ื›ื“ื™ ืœื”ื“ื’ื™ื ืืช ื”ืžืกืคืจื™ื.
00:59
Facebook has 1.2 billion users per month.
21
59332
3477
ืœืคื™ื™ืกื‘ื•ืง ื™ืฉ 1.2 ืžื™ืœื™ืืจื“ ืžืฉืชืžืฉื™ื ืœื—ื•ื“ืฉ.
01:02
So half the Earth's Internet population
22
62809
1930
ืื– ื—ืฆื™ ืžืื•ื›ืœื•ืกื™ืช ื”ืื™ื ื˜ืจื ื˜ ื”ืขื•ืœืžื™ืช
01:04
is using Facebook.
23
64739
1653
ืžืฉืชืžืฉืช ื‘ืคื™ื™ืกื‘ื•ืง.
01:06
They are a site, along with others,
24
66392
1932
ื”ื ืืชืจ, ื™ื—ื“ ืขื ืื—ืจื™ื,
01:08
that has allowed people to create an online persona
25
68324
3219
ืฉืื™ืคืฉืจ ืœืื ืฉื™ื ืœื™ืฆื•ืจ ื™ืฉื•ืช ืื™ื ื˜ืจื ื˜ื™ืช
01:11
with very little technical skill,
26
71543
1782
ืขื ืžืขื˜ ืžืื•ื“ ื›ื™ืฉื•ืจื™ื ื˜ื›ื ื™ื™ื,
01:13
and people responded by putting huge amounts
27
73325
2476
ื•ืื ืฉื™ื ื”ื’ื™ื‘ื• ื‘ืœื”ืขืœื•ืช ื›ืžื•ื™ื•ืช ืขืฆื•ืžื•ืช
01:15
of personal data online.
28
75801
1983
ืฉืœ ืžื™ื“ืข ืื™ืฉื™ ืœืจืฉืช.
01:17
So the result is that we have behavioral,
29
77784
2543
ืื– ื”ืชื•ืฆืื” ื”ื™ื ืฉื™ืฉ ืœื ื• ืžื™ื“ืข
01:20
preference, demographic data
30
80327
1986
ื”ืชื ื”ื’ื•ืชื™, ื“ืžื•ื’ืจืคื™ ื•ื”ืขื“ืคื•ืช
01:22
for hundreds of millions of people,
31
82313
2101
ืฉืœ ืžืื•ืช ืžืœื™ื•ื ื™ ืื ืฉื™ื,
01:24
which is unprecedented in history.
32
84414
2026
ืฉื–ื” ืžืขื•ืœื ืœื ืงืจื” ื‘ื”ืกื˜ื•ืจื™ื”.
01:26
And as a computer scientist, what this means is that
33
86440
2560
ื•ื›ืžื“ืขื ื™ืช ืžื—ืฉื‘, ืžื” ืฉื–ื” ืื•ืžืจ
01:29
I've been able to build models
34
89000
1664
ื–ื” ืฉื”ื™ื™ืชื™ ืžืกื•ื’ืœืช ืœื‘ื ื•ืช ืžื•ื“ืœื™ื
01:30
that can predict all sorts of hidden attributes
35
90664
2322
ืฉื™ื›ื•ืœื™ื ืœืฆืคื•ืช ื›ืœ ืžื™ื ื™ ืชื›ื•ื ื•ืช ื—ื‘ื•ื™ื•ืช
01:32
for all of you that you don't even know
36
92986
2284
ืขื‘ื•ืจ ื›ื•ืœื›ื ืฉืืชื ืืคื™ืœื• ืœื ื™ื“ืขืชื
01:35
you're sharing information about.
37
95270
2202
ืฉืืชื ื—ื•ืœืงื™ื ืžื™ื“ืข ืขืœื™ื”ืŸ.
01:37
As scientists, we use that to help
38
97472
2382
ื›ืžื“ืขื ื™ื, ืื ื—ื ื• ืžืฉืชืžืฉื™ื ื‘ื–ื” ื›ื“ื™ ืœืขื–ื•ืจ
01:39
the way people interact online,
39
99854
2114
ืœืื ืฉื™ื ื‘ื“ืจืš ืฉื‘ื” ื”ื ืžืชืงืฉืจื™ื ื‘ืจืฉืช.
01:41
but there's less altruistic applications,
40
101968
2499
ืื‘ืœ ื™ืฉ ื™ืฉื•ืžื™ื ืคื—ื•ืช ืืœื˜ืจื•ืื™ืกื˜ื™ื,
01:44
and there's a problem in that users don't really
41
104467
2381
ื•ื™ืฉ ื‘ืขื™ื” ื‘ื–ื” ืฉืžืชืฉืžืฉื™ื ืœื ื‘ืืžืช
01:46
understand these techniques and how they work,
42
106848
2470
ืžื‘ื™ื ื™ื ืืช ื”ื˜ื›ื ื™ืงื•ืช ื•ืื™ืš ื”ืŸ ืขื•ื‘ื“ื•ืช,
01:49
and even if they did, they don't have a lot of control over it.
43
109318
3128
ื•ืืคื™ืœื• ืื ื”ื ื”ื‘ื™ื ื•, ืื™ืŸ ืœื”ื ื”ืจื‘ื” ืฉืœื™ื˜ื” ืขืœ ื–ื”.
01:52
So what I want to talk to you about today
44
112446
1490
ืื– ืžื” ืฉืื ื™ ืจื•ืฆื” ืœืกืคืจ ืœื›ื ื”ื™ื•ื
01:53
is some of these things that we're able to do,
45
113936
2702
ื–ื” ื—ืœืง ืžื”ื“ื‘ืจื™ื ืฉืื ื—ื ื• ื™ื›ื•ืœื™ื ืœืขืฉื•ืช,
01:56
and then give us some ideas of how we might go forward
46
116638
2763
ื•ืื– ืœืชืช ืœื›ื ื›ืžื” ืจืขื™ื•ื ื•ืช ืขืœ ืื™ืš ื ืžืฉื™ืš ืžืคื”
01:59
to move some control back into the hands of users.
47
119401
2769
ืœื”ื–ื™ื– ืงืฆืช ืฉืœื™ื˜ื” ื—ื–ืจื” ืœื™ื“ื™ื™ื ืฉืœ ื”ืžืฉืชืžืฉื™ื.
02:02
So this is Target, the company.
48
122170
1586
ืื– ื–ื• ื˜ืืจื’ื˜, ื”ื—ื‘ืจื”.
02:03
I didn't just put that logo
49
123756
1324
ืœื ืกืชื ืฉืžืชื™ ืืช ื”ืœื•ื’ื•
02:05
on this poor, pregnant woman's belly.
50
125080
2170
ืขืœ ื”ื‘ื˜ืŸ ืฉืœ ื”ืื™ืฉื” ื”ื”ืจื™ื•ื ื™ืช ื”ืžืกื›ื ื” ื”ื–ื•.
02:07
You may have seen this anecdote that was printed
51
127250
1840
ืืชื ืื•ืœื™ ืจืื™ืชื ืืช ื”ืื ืงื“ื•ื˜ื” ื”ื–ื• ืžื•ื“ืคืกืช
02:09
in Forbes magazine where Target
52
129090
2061
ื‘ืžื’ื–ื™ืŸ ืคื•ืจื‘ืก ื›ืฉื˜ืืจื’ื˜
02:11
sent a flyer to this 15-year-old girl
53
131151
2361
ืฉืœื—ื” ืขืœื•ืŸ ืœื‘ืช ื”15 ื”ื–ื•
02:13
with advertisements and coupons
54
133512
1710
ืขื ืคืจืกื•ืžื•ืช ื•ืงื•ืคื•ื ื™ื
02:15
for baby bottles and diapers and cribs
55
135222
2554
ืœื‘ืงื‘ื•ืงื™ ืชื™ื ื•ืงื•ืช ื•ื—ื™ืชื•ืœื™ื ื•ืขืจื™ืกื•ืช
02:17
two weeks before she told her parents
56
137776
1684
ืฉื‘ื•ืขื™ื™ื ืœืคื ื™ ืฉืืžืจื” ืœื”ื•ืจื™ื”
02:19
that she was pregnant.
57
139460
1864
ืฉื”ื™ื ื‘ื”ืจื™ื•ืŸ,
02:21
Yeah, the dad was really upset.
58
141324
2704
ื›ืŸ, ื”ืื‘ื ื”ื™ื” ืžืžืฉ ืขืฆื‘ื ื™.
02:24
He said, "How did Target figure out
59
144028
1716
ื”ื•ื ืืžืจ, "ืื™ืš ื˜ืืจื’ื˜ ื”ื‘ื™ื ื•
02:25
that this high school girl was pregnant
60
145744
1824
ืฉื ืขืจื” ื‘ืชื™ื›ื•ืŸ ื‘ื”ืจื™ื•ืŸ
02:27
before she told her parents?"
61
147568
1960
ืœืคื ื™ ืฉื”ื™ื ืืžืจื” ืœื”ื•ืจื™ื”?"
02:29
It turns out that they have the purchase history
62
149528
2621
ืžืกืชื‘ืจ ืฉื™ืฉ ืœื”ื ืืช ื”ืกื˜ื•ืจื™ืช ื”ืจื›ื™ืฉื•ืช
02:32
for hundreds of thousands of customers
63
152149
2301
ืœืžืื•ืช ืืœืคื™ ืœืงื•ื—ื•ืช
02:34
and they compute what they call a pregnancy score,
64
154450
2730
ื•ื”ื ืžื—ืฉื‘ื™ื ืืช ืžื” ืฉื”ื ืงื•ืจืื™ื ืœื• ืฆื™ื•ืŸ ื”ืจื™ื•ืŸ,
02:37
which is not just whether or not a woman's pregnant,
65
157180
2332
ืฉื–ื” ืœื ืจืง ืื ืื™ืฉื” ื‘ื”ืจื™ื•ืŸ,
02:39
but what her due date is.
66
159512
1730
ืืœื ืžืชื™ ื”ืชืืจื™ืš ื”ืžื™ื•ืขื“.
02:41
And they compute that
67
161242
1304
ื•ื”ื ืžื—ืฉื‘ื™ื ืืช ื–ื”
02:42
not by looking at the obvious things,
68
162546
1768
ืœื ืขืœ ื™ื“ื™ ื”ืกืชื›ืœื•ืช ืขืœ ื“ื‘ืจื™ื ื‘ืจื•ืจื™ื,
02:44
like, she's buying a crib or baby clothes,
69
164314
2512
ื›ืžื•, ื”ื™ื ืงื•ื ื” ืขืจื™ืกื” ืื• ื‘ื’ื“ื™ ืชื™ื ื•ืง,
02:46
but things like, she bought more vitamins
70
166826
2943
ืืœื ื“ื‘ืจื™ื ื›ืžื•, ื”ื™ื ืงื ืชื” ื™ื•ืชืจ ื•ื™ื˜ืžื™ื ื™ื
02:49
than she normally had,
71
169769
1717
ืžืฉื”ื™ื ื‘ื“ืจืš ื›ืœืœ ืงื•ื ื”,
02:51
or she bought a handbag
72
171486
1464
ืื• ืฉื”ื™ื ืงื ืชื” ืชื™ืง
02:52
that's big enough to hold diapers.
73
172950
1711
ืฉืžืกืคื™ืง ื’ื“ื•ืœ ืœื”ื—ื–ื™ืง ื—ื™ืชื•ืœื™ื.
02:54
And by themselves, those purchases don't seem
74
174661
1910
ื•ื‘ืขืฆืžืŸ, ืœืžืจื•ืช ืฉื”ืงื ื™ื•ืช ื”ืืœื” ืœื ื ืจืื•ืช
02:56
like they might reveal a lot,
75
176571
2469
ื›ืื™ืœื• ื”ืŸ ืžื’ืœื•ืช ื”ืจื‘ื”,
02:59
but it's a pattern of behavior that,
76
179040
1978
ืื‘ืœ ื–ื• ืชื‘ื ื™ืช ื”ืชื ื”ื’ื•ืช,
03:01
when you take it in the context of thousands of other people,
77
181018
3117
ืฉื›ืฉืืชื ืžื›ื ื™ืกื™ื ืืช ื–ื” ืœื”ืงืฉืจ ืฉืœ ืืœืคื™ ืื ืฉื™ื ืื—ืจื™ื,
03:04
starts to actually reveal some insights.
78
184135
2757
ืžืชื—ื™ืœื” ืœืžืขืฉื” ืœื’ืœื•ืช ืชื•ื‘ื ื•ืช.
03:06
So that's the kind of thing that we do
79
186892
1793
ืื– ื–ื” ืกื•ื’ ื”ื“ื‘ืจื™ื ืฉืื ื—ื ื• ืขื•ืฉื™ื
03:08
when we're predicting stuff about you on social media.
80
188685
2567
ื›ืฉืื ื—ื ื• ื—ื•ื–ื™ื ื“ื‘ืจื™ื ืขืœื™ื›ื ื‘ืžื“ื™ื” ื”ื—ื‘ืจืชื™ืช.
03:11
We're looking for little patterns of behavior that,
81
191252
2796
ืื ื—ื ื• ืžื—ืคืฉื™ื ืชื‘ื ื™ื•ืช ื–ืขื™ืจื•ืช ืฉืœ ื”ืชื ื”ื’ื•ืช,
03:14
when you detect them among millions of people,
82
194048
2682
ืฉื›ืฉืžื–ื”ื™ื ืื•ืชืŸ ื‘ื™ืŸ ืžืœื™ื•ื ื™ ืื ืฉื™ื,
03:16
lets us find out all kinds of things.
83
196730
2706
ื ื•ืชื ื•ืช ืœื ื• ืœืžืฆื•ื ื›ืœ ืžื™ื ื™ ื“ื‘ืจื™ื.
03:19
So in my lab and with colleagues,
84
199436
1747
ืื– ื‘ืžืขื‘ื“ื” ืฉืœื™ ื•ืขื ืฉื•ืชืคื™ื,
03:21
we've developed mechanisms where we can
85
201183
1777
ืื ื—ื ื• ืคื™ืชื—ื ื• ืžื›ืื ื™ื–ืžื™ื ืื™ืชื ืื ื—ื ื• ื™ื›ื•ืœื™ื
03:22
quite accurately predict things
86
202960
1560
ืœื—ื–ื•ืช ื“ื™ ื‘ืžื“ื•ื™ื™ืง ื“ื‘ืจื™ื
03:24
like your political preference,
87
204520
1725
ื›ืžื• ื”ืขื“ืคื•ืช ืคื•ืœื™ื˜ื™ื•ืช,
03:26
your personality score, gender, sexual orientation,
88
206245
3752
ืฆื™ื•ืŸ ื”ืื™ืฉื™ื•ืช ืฉืœื›ื, ืžื™ืŸ, ื”ืขื“ืคื•ืช ืžื™ื ื™ื•ืช,
03:29
religion, age, intelligence,
89
209997
2873
ื“ืช, ื’ื™ืœ, ืจืžืช ืื™ื ื˜ื™ืœื™ื’ื ืฆื™ื”,
03:32
along with things like
90
212870
1394
ื™ื—ื“ ืขื ื“ื‘ืจื™ื ื›ืžื•
03:34
how much you trust the people you know
91
214264
1937
ื›ืžื” ืืชื ื‘ื•ื˜ื—ื™ื ื‘ืื ืฉื™ื ืื•ืชื ืืชื ืžื›ื™ืจื™ื
03:36
and how strong those relationships are.
92
216201
1804
ื•ื›ืžื” ื—ื–ืงื™ื ื”ืงืฉืจื™ื ื”ืืœื”.
03:38
We can do all of this really well.
93
218005
1785
ืื ื—ื ื• ื™ื›ื•ืœื™ื ืœืขืฉื•ืช ืืช ื–ื” ืžืžืฉ ื˜ื•ื‘.
03:39
And again, it doesn't come from what you might
94
219790
2197
ื•ืฉื•ื‘, ื–ื” ืœื ืžื’ื™ืข ืžืžื” ืฉืืชื ืื•ืœื™
03:41
think of as obvious information.
95
221987
2102
ื—ื•ืฉื‘ื™ื ื›ืžื™ื“ืข ื‘ืจื•ืจ.
03:44
So my favorite example is from this study
96
224089
2281
ืื– ื”ื“ื•ื’ืžื” ื”ืื”ื•ื‘ื” ืขืœื™ ื”ื™ื ืžืžื—ืงืจ
03:46
that was published this year
97
226370
1240
ืฉืคื•ืจืกื ื”ืฉื ื”
03:47
in the Proceedings of the National Academies.
98
227610
1795
ื‘ืคืจืกื•ืžื™ื ืฉืœ ื”ืืงื“ืžื™ื” ื”ืœืื•ืžื™ืช.
03:49
If you Google this, you'll find it.
99
229405
1285
ืื ืชื’ื’ืœื• ืืช ื–ื”, ืืชื ืชืžืฆืื• ืื•ืชื•.
03:50
It's four pages, easy to read.
100
230690
1872
ื–ื” ืืจื‘ืขื” ื“ืคื™ื, ืงืœื™ื ืœืงืจื™ืื”.
03:52
And they looked at just people's Facebook likes,
101
232562
3003
ื•ื”ื ื”ืกืชื›ืœื• ืจืง ืขืœ ืœื™ื™ืงื™ื ืฉืœ ืื ืฉื™ื ื‘ืคื™ื™ืกื‘ื•ืง,
03:55
so just the things you like on Facebook,
102
235565
1920
ืื– ืจืง ื”ื“ื‘ืจื™ื ืฉืืชื ืื•ื”ื‘ื™ื ื‘ืคื™ื™ืกื‘ื•ืง,
03:57
and used that to predict all these attributes,
103
237485
2138
ื•ื”ืฉืชืžืฉื• ื‘ื–ื” ื›ื“ื™ ืœื—ื–ื•ืช ืืช ื›ืœ ื”ืชื›ื•ื ื•ืช ื”ืืœื”,
03:59
along with some other ones.
104
239623
1645
ื™ื—ื“ ืขื ื›ืžื” ืื—ืจื•ืช.
04:01
And in their paper they listed the five likes
105
241268
2961
ื•ื‘ืžืืžืจ ืฉืœื”ื ื”ื ืฆื™ื™ื ื• ืืช ื—ืžืฉืช ื”ืœื™ื™ืงื™ื
04:04
that were most indicative of high intelligence.
106
244229
2787
ืฉื”ื›ื™ ื”ืจืื• ืื™ื ื˜ื™ืœื™ื’ื ืฆื™ื” ื’ื‘ื•ื”ื”.
04:07
And among those was liking a page
107
247016
2324
ื•ื‘ื™ื ื”ื ื”ื™ื” ืœื™ื™ืง ืœื“ืฃ
04:09
for curly fries. (Laughter)
108
249340
1905
ืœืฆ'ื™ืคืกื™ื ืžืงื•ืจื–ืœื™ื. (ืฆื—ื•ืง)
04:11
Curly fries are delicious,
109
251245
2093
ืฆ'ื™ืคืกื™ื ืžืงื•ืจื–ืœื™ื ื”ื ืžืžืฉ ื˜ืขื™ืžื™ื,
04:13
but liking them does not necessarily mean
110
253338
2530
ืื‘ืœ ืื”ื‘ื” ืฉืœื”ื ืœื ื‘ื”ื›ืจื— ืื•ืžืจืช
04:15
that you're smarter than the average person.
111
255868
2080
ืฉืืชื ืื“ื ื—ื›ื ืžื”ืžืžื•ืฆืข.
04:17
So how is it that one of the strongest indicators
112
257948
3207
ืื– ืื™ืš ืื—ื“ ื”ืžื“ื“ื™ื ื”ื—ื–ืงื™ื ื‘ื™ื•ืชืจ
04:21
of your intelligence
113
261155
1570
ืœืื™ื ื˜ืœื™ื’ื ืฆื™ื” ืฉืœื›ื
04:22
is liking this page
114
262725
1447
ื”ื•ื ืื”ื‘ืช ื”ื“ืฃ ื”ื–ื”
04:24
when the content is totally irrelevant
115
264172
2252
ื›ืฉื”ืชื•ื›ืŸ ืœื—ืœื•ื˜ื™ืŸ ืœื ืจืœื•ื•ื ื˜ื™
04:26
to the attribute that's being predicted?
116
266424
2527
ืœืชื›ื•ื ื•ืช ืฉื ืฆืคื•ืช?
04:28
And it turns out that we have to look at
117
268951
1584
ื•ืžืกืชื‘ืจ ืฉืื ื—ื ื• ืฆืจื™ื›ื™ื ืœื”ื‘ื™ื˜
04:30
a whole bunch of underlying theories
118
270535
1618
ื‘ืงื‘ื•ืฆื” ืฉืœืžื” ืฉืœ ืชืื•ืจื™ื•ืช
04:32
to see why we're able to do this.
119
272153
2569
ื›ื“ื™ ืœืจืื•ืช ืœืžื” ืื ื—ื ื• ืžืกื•ื’ืœื™ื ืœืขืฉื•ืช ืืช ื–ื”.
04:34
One of them is a sociological theory called homophily,
120
274722
2913
ืื—ืช ืžื”ืŸ ื”ื™ื ืชืื•ืจื™ื” ืกื•ืฆื™ื•ืœื•ื’ื™ืช ืฉื ืงืจืืช ื”ื•ืžื•ืคื™ืœื™ื”,
04:37
which basically says people are friends with people like them.
121
277635
3092
ืฉืื•ืžืจืช ื‘ืขื™ืงืจื•ืŸ ืฉืื ืฉื™ื ื—ื‘ืจื™ื ืขื ืื ืฉื™ื ืฉื“ื•ืžื™ื ืœื”ื.
04:40
So if you're smart, you tend to be friends with smart people,
122
280727
2014
ืื– ืื ืืชื” ื—ื›ื, ืืชื” ื ื•ื˜ื” ืœื”ื™ื•ืช ื—ื‘ืจ ืฉืœ ืื ืฉื™ื ื—ื›ืžื™ื,
04:42
and if you're young, you tend to be friends with young people,
123
282741
2630
ื•ืื ืืชื” ื”ืฆืขื™ืจ, ืืชื” ื ื•ื˜ื” ืœื”ื™ื•ืช ื—ื‘ืจ ืฉืœ ืื ืฉื™ื ืฆืขื™ืจื™ื,
04:45
and this is well established
124
285371
1627
ื•ื–ื” ืžื•ื›ื— ื”ื™ื˜ื‘
04:46
for hundreds of years.
125
286998
1745
ืžืื•ืช ืฉื ื™ื.
04:48
We also know a lot
126
288743
1232
ืื ื—ื ื• ื’ื ื™ื•ื“ืขื™ื ื”ืจื‘ื”
04:49
about how information spreads through networks.
127
289975
2550
ืขืœ ืื™ืš ืžื™ื“ืข ืžืชืคืฉื˜ ื‘ืจืฉืชื•ืช.
04:52
It turns out things like viral videos
128
292525
1754
ืžืกืชื‘ืจ ืฉื“ื‘ืจื™ื ื›ืžื• ืกืจื˜ื•ื ื™ื ื•ื™ืจืืœื™ื™ื
04:54
or Facebook likes or other information
129
294279
2406
ืื• ืœื™ื™ืงื™ื ืฉืœ ืคื™ื™ืกื‘ื•ืง ืื• ืžื™ื“ืข ืื—ืจ
04:56
spreads in exactly the same way
130
296685
1888
ืžืชืคืฉื˜ื™ื ื‘ื“ื™ื•ืง ื‘ืื•ืชื” ื“ืจืš
04:58
that diseases spread through social networks.
131
298573
2454
ืฉืžื—ืœื•ืช ืžืชืคืฉื˜ื•ืช ื‘ืจืฉืชื•ืช ื—ื‘ืจืชื™ื•ืช.
05:01
So this is something we've studied for a long time.
132
301027
1791
ืื– ื–ื” ืžืฉื”ื• ืฉื—ืงืจื ื• ื”ืจื‘ื” ื–ืžืŸ.
05:02
We have good models of it.
133
302818
1576
ื™ืฉ ืœื ื• ืžื•ื“ืœื™ื ื˜ื•ื‘ื™ื ืฉืœ ื–ื”.
05:04
And so you can put those things together
134
304394
2157
ื•ื›ืš ืืชื ื™ื›ื•ืœื™ื ืœื—ื‘ืจ ืืช ื”ื“ื‘ืจื™ื ื”ืืœื”
05:06
and start seeing why things like this happen.
135
306551
3088
ื•ืœื”ืชื—ื™ืœ ืœืจืื•ืช ืœืžื” ื“ื‘ืจื™ื ื›ืืœื” ืงื•ืจื™ื.
05:09
So if I were to give you a hypothesis,
136
309639
1814
ืื– ืื ื”ื™ื™ืชื™ ื ื•ืชื ืช ืœื›ื ื”ืฉืขืจื”,
05:11
it would be that a smart guy started this page,
137
311453
3227
ื”ื™ื ืชื”ื™ื” ืฉืื“ื ื—ื›ื ื”ืชื—ื™ืœ ืืช ื”ื“ืฃ ื”ื–ื”,
05:14
or maybe one of the first people who liked it
138
314680
1939
ืื• ืื•ืœื™ ืื—ื“ ื”ืื ืฉื™ื ื”ืจืืฉื•ื ื™ื ืฉืื”ื‘ื• ืื•ืชื•
05:16
would have scored high on that test.
139
316619
1736
ื”ื™ื” ืžืงื‘ืœ ืฆื™ื•ืŸ ื’ื‘ื•ื” ื‘ืžื‘ื—ืŸ ื”ื–ื”.
05:18
And they liked it, and their friends saw it,
140
318355
2288
ื•ื”ื ืื”ื‘ื• ืื•ืชื•, ื•ื—ื‘ืจื™ื ืฉืœื”ื ืจืื• ืื•ืชื•,
05:20
and by homophily, we know that he probably had smart friends,
141
320643
3122
ื•ื“ืจืš ื”ื”ื•ืžื•ืคื™ืœื™ื”, ืื ื—ื ื• ื™ื•ื“ืขื™ื ืฉื›ื ืจืื” ื”ื™ื• ืœื• ื—ื‘ืจื™ื ื—ื›ืžื™ื,
05:23
and so it spread to them, and some of them liked it,
142
323765
3056
ืื– ื–ื” ื”ืชืคืฉื˜ ืืœื™ื”ื, ื•ื›ืžื” ืžื”ื ืื”ื‘ื• ืืช ื–ื”,
05:26
and they had smart friends,
143
326821
1189
ื•ืœื”ื ื”ื™ื• ื—ื‘ืจื™ื ื—ื›ืžื™ื,
05:28
and so it spread to them,
144
328010
807
05:28
and so it propagated through the network
145
328817
1973
ื•ื›ืš ื–ื” ื”ืชืคืฉื˜ ืืœื™ื”ื,
ื•ืื– ื–ื” ื—ืœื—ืœ ื“ืจืš ื”ืจืฉืช
05:30
to a host of smart people,
146
330790
2569
ืœื”ืจื‘ื” ืื ืฉื™ื ื—ื›ืžื™ื,
05:33
so that by the end, the action
147
333359
2056
ืื– ื‘ืกื•ืฃ, ื”ืคืขื•ืœื”
05:35
of liking the curly fries page
148
335415
2544
ืฉืœ ืื”ื‘ืช ื“ืฃ ื”ืฆ'ื™ืคืกื™ื ื”ืžืงื•ืจื–ืœื™ื
05:37
is indicative of high intelligence,
149
337959
1615
ืžืขื™ื“ื” ืฉืœ ืื™ื ื˜ืœื™ื’ื ืฆื™ื” ื’ื‘ื•ื”ื”,
05:39
not because of the content,
150
339574
1803
ืœื ื‘ื’ืœืœ ื”ืชื•ื›ืŸ,
05:41
but because the actual action of liking
151
341377
2522
ืืœื ื‘ื’ืœืœ ื”ืคืขื•ืœื” ืขืฆืžื” ืฉืœ ืื”ื‘ื”
05:43
reflects back the common attributes
152
343899
1900
ืžืฉืงืคืช ืืช ื”ืชื›ื•ื ื•ืช ื”ืžืฉื•ืชืคื•ืช
05:45
of other people who have done it.
153
345799
2468
ืฉืœ ืื ืฉื™ื ืื—ืจื™ื ืฉืขืฉื• ืืช ื–ื”.
05:48
So this is pretty complicated stuff, right?
154
348267
2897
ืื– ืืœื” ื“ื‘ืจื™ื ื“ื™ ืžืกื•ื‘ื›ื™ื, ื ื›ื•ืŸ?
05:51
It's a hard thing to sit down and explain
155
351164
2199
ื–ื” ืงืฉื” ืœืฉื‘ืช ืœื”ืกื‘ื™ืจ
05:53
to an average user, and even if you do,
156
353363
2848
ืœืžืฉืชืžืฉ ืžืžื•ืฆืข, ื•ืืคื™ืœื• ืื ืืชื ืขื•ืฉื™ื ื–ืืช,
05:56
what can the average user do about it?
157
356211
2188
ืžื” ื”ืžืฉืชืžืฉ ื”ืžืžื•ืฆืข ื™ื›ื•ืœ ืœืขืฉื•ืช ื‘ื ื•ื’ืข ืœื–ื”?
05:58
How do you know that you've liked something
158
358399
2048
ืื™ืš ืืชื ื™ื•ื“ืขื™ื ืฉืื”ื‘ืชื ืžืฉื”ื•
06:00
that indicates a trait for you
159
360447
1492
ืฉืžืฉืงืฃ ืชื›ื•ื ื” ืฉืœื›ื
06:01
that's totally irrelevant to the content of what you've liked?
160
361939
3545
ืฉืœื’ืžืจื™ ืœื ืจืœื•ื•ื ื˜ื™ืช ืœืชื•ื›ืŸ ืฉืœ ืžื” ืฉืื”ื‘ืชื?
06:05
There's a lot of power that users don't have
161
365484
2546
ื™ืฉ ื”ืจื‘ื” ื›ื•ื— ืฉืื™ืŸ ืœืžืฉืชืžืฉื™ื
06:08
to control how this data is used.
162
368030
2230
ื›ื“ื™ ืœืฉืœื•ื˜ ื‘ืื™ืš ืžืฉืชืžืฉื™ื ื‘ืžื™ื“ืข ื”ื–ื”.
06:10
And I see that as a real problem going forward.
163
370260
3112
ื•ืื ื™ ืจื•ืื” ื‘ื–ื” ื‘ืขื™ื” ืืžื™ืชื™ืช ื‘ืขืชื™ื“.
06:13
So I think there's a couple paths
164
373372
1977
ืื– ืื ื™ ื—ื•ืฉื‘ืช ืฉื™ืฉ ื›ืžื” ื›ื™ื•ื•ื ื™ื
06:15
that we want to look at
165
375349
1001
ืฉื ืจืฆื” ืœื‘ื—ื•ืŸ
06:16
if we want to give users some control
166
376350
1910
ืื ืื ื—ื ื• ืจื•ืฆื™ื ืœืชืช ืœืžืฉืชืžืฉื™ื ืžืขื˜ ืฉืœื™ื˜ื”
06:18
over how this data is used,
167
378260
1740
ืขืœ ืื™ืš ืžืฉืชืžืฉื™ื ื‘ืžื™ื“ืข ื”ื–ื”,
06:20
because it's not always going to be used
168
380000
1940
ืžืคื ื™ ืฉืœื ืชืžื™ื“ ื”ื•ื ื™ื”ื™ื” ื‘ืฉื™ืžื•ืฉ
06:21
for their benefit.
169
381940
1381
ืœืชื•ืขืœืชื.
06:23
An example I often give is that,
170
383321
1422
ื“ื•ื’ืžื” ืฉืื ื™ ื ื•ืชื ืช ืœื–ื” ื”ืจื‘ื” ื”ื™ื,
06:24
if I ever get bored being a professor,
171
384743
1646
ืฉืื ืื™ ืคืขื ืืฉืชืขืžื ืœื”ื™ื•ืช ืคืจื•ืคืกื•ืจื™ืช,
06:26
I'm going to go start a company
172
386389
1653
ืื ื™ ืืงื™ื ื—ื‘ืจื”
06:28
that predicts all of these attributes
173
388042
1454
ืฉื—ื•ื–ื” ืืช ื›ืœ ื”ืชื›ื•ื ื•ืช ื”ืืœื”
06:29
and things like how well you work in teams
174
389496
1602
ื•ื“ื‘ืจื™ื ื›ืžื• ื›ืžื” ื˜ื•ื‘ ืืชื ืขื•ื‘ื“ื™ื ื‘ืฆื•ื•ืช
06:31
and if you're a drug user, if you're an alcoholic.
175
391098
2671
ื•ืื ืืชื ืžืฉืชืžืฉื™ื ื‘ืกืžื™ื, ืื ืืชื ืืœื›ื•ื”ื•ืœื™ืกื˜ื™ื.
06:33
We know how to predict all that.
176
393769
1440
ืื ื—ื ื• ื™ื•ื“ืขื™ื ืื™ืš ืœื—ื–ื•ืช ืืช ื›ืœ ื–ื”.
06:35
And I'm going to sell reports
177
395209
1761
ื•ืื ื™ ืืžื›ื•ืจ ื“ื•ื—ื•ืช
06:36
to H.R. companies and big businesses
178
396970
2100
ืœื—ื‘ืจื•ืช ื›ื•ื— ืื“ื ื•ืขืกืงื™ื ื’ื“ื•ืœื™ื
06:39
that want to hire you.
179
399070
2273
ืฉืจื•ืฆื™ื ืœื”ืขืกื™ืง ืืชื›ื.
06:41
We totally can do that now.
180
401343
1177
ืื ื—ื ื• ืœื’ืžืจื™ ื™ื›ื•ืœื™ื ืœืขืฉื•ืช ืืช ื–ื” ืขื›ืฉื™ื•.
06:42
I could start that business tomorrow,
181
402520
1788
ื”ื™ื™ืชื™ ื™ื›ื•ืœื” ืœื”ืชื—ื™ืœ ืืช ื”ืขืกืง ื”ื–ื” ืžื—ืจ,
06:44
and you would have absolutely no control
182
404308
2052
ื•ืœื ื”ื™ืชื” ืœื›ื ืฉืœื™ื˜ื” ื‘ื›ืœืœ
06:46
over me using your data like that.
183
406360
2138
ืขืœ ื”ืฉื™ืžื•ืฉ ืฉืœื™ ื‘ืžื™ื“ืข ืฉืœื›ื ื›ืš.
06:48
That seems to me to be a problem.
184
408498
2292
ื–ื• ื ืจืื™ืช ืœื™ ื‘ืขื™ื” ื’ื“ื•ืœื”.
06:50
So one of the paths we can go down
185
410790
1910
ืื– ืื—ืช ื”ื“ืจื›ื™ื ืฉืื ื ื—ื• ื™ื›ื•ืœื™ื ืœื‘ื—ื•ืจ ื‘ื”
06:52
is the policy and law path.
186
412700
2032
ื”ื™ื ื”ืžื“ื™ื ื™ื•ืช ื•ื”ื—ื•ืง.
06:54
And in some respects, I think that that would be most effective,
187
414732
3046
ื•ื‘ื›ืžื” ื”ื‘ื˜ื™ื, ืื ื™ ื—ื•ืฉื‘ ืฉื–ื” ื™ื”ื™ื” ื”ื›ื™ ืืคืงื˜ื™ื‘ื™,
06:57
but the problem is we'd actually have to do it.
188
417778
2756
ืื‘ืœ ื”ื‘ืขื™ื” ื”ื™ื ืฉืœืžืขืฉื” ื ืฆื˜ืจืš ืœืขืฉื•ืช ืืช ื–ื”.
07:00
Observing our political process in action
189
420534
2780
ื‘ื”ืกืชื›ืœื•ืช ืขืœ ื”ืชื”ืœื™ืš ื”ืคื•ืœื™ื˜ื™ ืฉืœื ื• ื‘ืคืขื•ืœื”
07:03
makes me think it's highly unlikely
190
423314
2379
ื’ื•ืจื ืœื™ ืœื—ืฉื•ื‘ ืฉื–ื” ืžืื•ื“ ืœื ืกื‘ื™ืจ
07:05
that we're going to get a bunch of representatives
191
425693
1597
ืฉื ืฉื™ื’ ื›ืžื” ื ื‘ื—ืจื™ื
07:07
to sit down, learn about this,
192
427290
1986
ืœืฉื‘ืช, ืœืœืžื•ื“ ืืช ื”ื ื•ืฉื,
07:09
and then enact sweeping changes
193
429276
2106
ื•ืื– ืœื”ื—ื™ืœ ืฉื™ื ื•ื™ื™ื ืžืงื™ืคื™ื
07:11
to intellectual property law in the U.S.
194
431382
2157
ืขืœ ื—ื•ืงื™ ืจื›ื•ืฉ ืจืขื™ื•ื ื™ ื‘ืืจื”"ื‘.
07:13
so users control their data.
195
433539
2461
ื›ืš ืฉืžืฉืชืžืฉื™ื ื™ืฉืœื˜ื• ื‘ืžื™ื“ืข ืฉืœื”ื.
07:16
We could go the policy route,
196
436000
1304
ื ื•ื›ืœ ืœืœื›ืช ื‘ื“ืจืš ื”ืžื“ื™ื ื™ื•ืช,
07:17
where social media companies say,
197
437304
1479
ืฉื ื—ื‘ืจื•ืช ืžื“ื™ื” ื—ื‘ืจืชื™ืช ืื•ืžืจื•ืช,
07:18
you know what? You own your data.
198
438783
1402
ืืชื ื™ื•ื“ืขื™ื ืžื”? ื”ืžื™ื“ืข ื‘ื‘ืขืœื•ืชื›ื.
07:20
You have total control over how it's used.
199
440185
2489
ื™ืฉ ืœื›ื ืฉืœื™ื˜ื” ืžืœืื” ื‘ืฉื™ืžื•ืฉ ื‘ื•.
07:22
The problem is that the revenue models
200
442674
1848
ื”ื‘ืขื™ื” ื”ื™ื ืฉืžื•ื“ืœื™ ื”ืจื•ื•ื—ื™ื•ืช
07:24
for most social media companies
201
444522
1724
ืœืจื•ื‘ ื—ื‘ืจื•ืช ื”ืžื“ื™ื” ื”ื—ื‘ืจืชื™ืช
07:26
rely on sharing or exploiting users' data in some way.
202
446246
4031
ื ืกืžื›ื™ื ืขืœ ื ื™ืฆื•ืœ ื”ืžื™ื“ืข ืฉืœ ื”ืžืฉืชืžืฉ ื‘ื“ืจืš ื›ืœืฉื”ื™.
07:30
It's sometimes said of Facebook that the users
203
450277
1833
ืœืคืขืžื™ื ื ืืžืจ ืขืœ ืคื™ื™ืกื‘ื•ืง ืฉื”ืžืฉืชืžืฉื™ื
07:32
aren't the customer, they're the product.
204
452110
2528
ื”ื ืœื ืœืงื•ื—ื•ืช, ื”ื ื”ืžื•ืฆืจ.
07:34
And so how do you get a company
205
454638
2714
ื•ื›ืš ืื™ืš ืืชื ื’ื•ืจืžื™ื ืœื—ื‘ืจื”
07:37
to cede control of their main asset
206
457352
2558
ืœื•ื•ืชืจ ืขืœ ืฉืœื™ื˜ื” ื‘ื ื›ืก ื”ืขื™ืงืจื™ ืฉืœื”ื
07:39
back to the users?
207
459910
1249
ื‘ื—ื–ืจื” ืœืžืฉืชืžืฉื™ื?
07:41
It's possible, but I don't think it's something
208
461159
1701
ื–ื” ืืคืฉืจื™, ืื‘ืœ ืื ื™ ืœื ื—ื•ืฉื‘ ืฉื–ื” ืžืฉื”ื•
07:42
that we're going to see change quickly.
209
462860
2320
ืฉืื ื—ื ื• ื ืจืื” ืžืฉืชื ื” ื‘ืžื”ื™ืจื•ืช.
07:45
So I think the other path
210
465180
1500
ืื– ืื ื™ ื—ื•ืฉื‘ืช ืฉื”ื“ืจืš ื”ื ื•ืกืคืช
07:46
that we can go down that's going to be more effective
211
466680
2288
ืฉืื ื—ื ื• ื™ื›ื•ืœื™ื ืœืœื›ืช ื‘ื” ืฉืชื”ื™ื” ื™ื•ืชืจ ื™ืขื™ืœื”
07:48
is one of more science.
212
468968
1508
ื”ื™ื ื–ื• ืฉืœ ื™ื•ืชืจ ืžื“ืข.
07:50
It's doing science that allowed us to develop
213
470476
2510
ืœืขืฉื•ืช ืžื“ืข ื–ื” ืžื” ืฉืืคืฉืจ ืœื ื• ืœืคืชื—
07:52
all these mechanisms for computing
214
472986
1750
ืืช ื›ืœ ื”ืžื ื’ื ื•ื ื™ื ืœืžื—ืฉื•ื‘
07:54
this personal data in the first place.
215
474736
2052
ื”ืžื™ื“ืข ื”ืื™ืฉื™ ืžืจืืฉ.
07:56
And it's actually very similar research
216
476788
2106
ื•ื–ื” ืœืžืขืฉื” ืžื—ืงืจ ืžืื•ื“ ื“ื•ืžื”
07:58
that we'd have to do
217
478894
1438
ืฉืื ื—ื ื• ืฆืจื™ื›ื™ื ืœืขืฉื•ืช
08:00
if we want to develop mechanisms
218
480332
2386
ืื ื ืจืฆื” ืœืคืชื— ืžื ื’ื ื•ื ื™ื
08:02
that can say to a user,
219
482718
1421
ืฉื™ื›ื•ืœื™ื ืœื”ื’ื™ื“ ืœืžืฉืชืžืฉ,
08:04
"Here's the risk of that action you just took."
220
484139
2229
"ื”ื ื” ื”ืกื™ื›ื•ื ื™ื ืฉืœ ื”ืคืขื•ืœื” ื”ื–ื• ืฉืขืฉื™ืช."
08:06
By liking that Facebook page,
221
486368
2080
ืขืœ ื™ื“ื™ ืœื—ื™ืฆืช ืœื™ื™ืง ืขืœ ื“ืฃ ืคื™ื™ืกื‘ื•ืง ืžืกื•ื™ื™ื,
08:08
or by sharing this piece of personal information,
222
488448
2535
ืื• ืขืœื™ ื™ื“ื™ ืฉื™ืชื•ืฃ ืคื™ืกื” ื–ื• ืฉืœ ืžื™ื“ืข ืื™ืฉื™,
08:10
you've now improved my ability
223
490983
1502
ืฉื™ืคืจืชื ืขื›ืฉื™ื• ืืช ื”ื™ื›ื•ืœืช ืฉืœื™
08:12
to predict whether or not you're using drugs
224
492485
2086
ืœื—ื–ื•ืช ืื ืืชื ืžืฉืชืžืฉื™ื ื‘ืกืžื™ื
08:14
or whether or not you get along well in the workplace.
225
494571
2862
ืื• ืื ืืชื ืžืกืชื“ืจื™ื ื‘ืžืงื•ื ื”ืขื‘ื•ื“ื”.
08:17
And that, I think, can affect whether or not
226
497433
1848
ื•ื–ื”, ืื ื™ ื—ื•ืฉื‘ืช, ื™ื›ื•ืœ ืœื”ืฉืคื™ืข ืขืœ ืื
08:19
people want to share something,
227
499281
1510
ืื ืฉื™ื ื™ืจืฆื• ืœื—ืœื•ืง ืžืฉื”ื•,
08:20
keep it private, or just keep it offline altogether.
228
500791
3239
ืœืฉืžื•ืจ ืขืœ ื–ื” ืคืจื˜ื™, ืื• ืคืฉื•ื˜ ืœืฉืžื•ืจ ืขืœ ื–ื” ืžื—ื•ืฅ ืœืจืฉืช.
08:24
We can also look at things like
229
504030
1563
ืื ื—ื ื• ื™ื›ื•ืœื™ื ืœื”ื‘ื™ื˜ ื’ื ื‘ื“ื‘ืจื™ื
08:25
allowing people to encrypt data that they upload,
230
505593
2728
ื›ืžื• ืœืืคืฉืจ ืœืื ืฉื™ื ืœื”ืฆืคื™ืŸ ืืช ื”ืžื™ื“ืข ืฉื”ื ืžืขืœื™ื,
08:28
so it's kind of invisible and worthless
231
508321
1855
ืื– ื”ื•ื ืกื•ื’ ืฉืœ ื‘ืœืชื™ ื ืจืื” ื•ื—ืกืจ ืชื•ืขืœืช
08:30
to sites like Facebook
232
510176
1431
ืœืืชืจื™ื ื›ืžื• ืคื™ื™ืกื‘ื•ืง
08:31
or third party services that access it,
233
511607
2629
ืื• ืฉืจื•ืชื™ ืฆื“ ืฉืœื™ืฉื™ ืฉื ื™ื’ืฉื™ื ืืœื™ื•,
08:34
but that select users who the person who posted it
234
514236
3247
ืื‘ืœ ืœืžืฉืชืžืฉื™ื ื”ื ื‘ื—ืจื™ื ืฉื”ืื“ื ืฉื”ืขืœื” ืืช ื–ื”
08:37
want to see it have access to see it.
235
517483
2670
ืจื•ืฆื” ืฉื™ืจืื• ืืช ื–ื”, ืชื”ื™ื” ื’ื™ืฉื” ืืœื™ื•.
08:40
This is all super exciting research
236
520153
2166
ื›ืœ ื–ื” ืžื—ืงืจ ืกื•ืคืจ ืžืจื’ืฉ
08:42
from an intellectual perspective,
237
522319
1620
ืžื ืงื•ื“ืช ืžื‘ื˜ ืื™ื ื˜ืœืงื˜ื•ืืœื™ืช,
08:43
and so scientists are going to be willing to do it.
238
523939
1859
ื•ื›ืš ืžื“ืขื ื™ื ื™ื”ื™ื• ืžื•ื›ื ื™ื ืœืขืฉื•ืช ืืช ื–ื”.
08:45
So that gives us an advantage over the law side.
239
525798
3610
ืื– ื–ื” ื ื•ืชืŸ ืœื ื• ื™ืชืจื•ืŸ ืขืœ ืืคืฉืจื•ืช ื”ื—ื•ืง.
08:49
One of the problems that people bring up
240
529408
1725
ืื—ืช ื”ื‘ืขื™ื•ืช ืฉืื ืฉื™ื ืžืขืœื™ื
08:51
when I talk about this is, they say,
241
531133
1595
ื›ืฉืื ื™ ืžื“ื‘ืจืช ืขืœ ื–ื”, ื”ื ืื•ืžืจื™ื,
08:52
you know, if people start keeping all this data private,
242
532728
2646
ืืช ื™ื•ื“ืขืช, ืื ืื ืฉื™ื ืžืชื—ื™ืœื™ื ืœืฉืžื•ืจ ืขืœ ื›ืœ ื”ืžื™ื“ืข ืคืจื˜ื™,
08:55
all those methods that you've been developing
243
535374
2113
ื›ืœ ื”ืฉื™ื˜ื•ืช ื”ืืœื” ืฉืคื™ืชื—ืชื
08:57
to predict their traits are going to fail.
244
537487
2653
ืœื—ื–ื•ืช ืืช ื”ืชื›ื•ื ื•ืช ืฉืœื”ื ื™ื›ืฉืœื•.
09:00
And I say, absolutely, and for me, that's success,
245
540140
3520
ื•ืื ื™ ืื•ืžืจืช, ื‘ื”ื—ืœื˜, ื•ื‘ืฉื‘ื™ืœื™, ื–ื• ื”ืฆืœื—ื”,
09:03
because as a scientist,
246
543660
1786
ืžืคื ื™ ืฉื›ืžื“ืขื ื™ืช,
09:05
my goal is not to infer information about users,
247
545446
3688
ื”ืžื˜ืจื” ืฉืœื™ ื”ื™ื ืœื ืœื”ืกื™ืง ืžื™ื“ืข ืขืœ ืžืฉืชืžืฉื™ื,
09:09
it's to improve the way people interact online.
248
549134
2767
ื–ื” ื›ื“ื™ ืœืฉืคืจ ืืช ื”ื“ืจืš ื‘ื” ืื ืฉื™ื ืžืชืงืฉืจื™ื ื‘ืจืฉืช.
09:11
And sometimes that involves inferring things about them,
249
551901
3218
ื•ืœืคืขืžื™ื ื–ื” ื“ื•ืจืฉ ืœื”ืกื™ืง ื“ื‘ืจื™ื ืขืœื™ื”ื,
09:15
but if users don't want me to use that data,
250
555119
3022
ืื‘ืœ ืื ืžืฉืชืžืฉื™ื ืœื ืจื•ืฆื™ื ืฉืื ื™ ืืฉืชืžืฉ ื‘ืžื™ื“ืข ื”ื–ื”,
09:18
I think they should have the right to do that.
251
558141
2038
ืื ื™ ื—ื•ืฉื‘ืช ืฉืฆืจื™ื›ื” ืœื”ื™ื•ืช ืœื”ื ื”ื–ื›ื•ืช ืœืขืฉื•ืช ืืช ื–ื”.
09:20
I want users to be informed and consenting
252
560179
2651
ืื ื™ ืจื•ืฆื” ืฉืžืฉืชืžืฉื™ื ื™ื”ื™ื• ืžื™ื•ื“ืขื™ื ื•ืžืกื›ื™ืžื™ื
09:22
users of the tools that we develop.
253
562830
2112
ืœื›ืœื™ื ืฉืื ื—ื ื• ืžืคืชื—ื™ื.
09:24
And so I think encouraging this kind of science
254
564942
2952
ืื– ืื ื™ ื—ื•ืฉื‘ืช ืฉืœืขื•ื“ื“ ืกื•ื’ ื–ื” ืฉืœ ืžื“ืข
09:27
and supporting researchers
255
567894
1346
ื•ืœืชืžื•ืš ื‘ื—ื•ืงืจื™ื
09:29
who want to cede some of that control back to users
256
569240
3023
ืฉืจื•ืฆื™ื ืœื”ื—ื–ื™ืจ ื—ืœืง ืžื”ืฉืœื™ื˜ื” ืœืžืฉืชืžืฉื™ื
09:32
and away from the social media companies
257
572263
2311
ื•ืœืงื—ืช ืื•ืชื• ืžื—ื‘ืจื•ืช ื”ืžื“ื™ื” ื”ื—ื‘ืจืชื™ืช
09:34
means that going forward, as these tools evolve
258
574574
2671
ืื•ืžืจ ืฉื›ืฉื ืžืฉื™ืš ื”ืœืื”, ื›ืฉื”ื›ืœื™ื ื”ืืœื” ื™ืชืคืชื—ื•
09:37
and advance,
259
577245
1476
ื•ื™ืชืงื“ืžื•,
09:38
means that we're going to have an educated
260
578721
1414
ืื•ืžืจ ืฉื™ื”ื™ื” ืœื ื• ื‘ืกื™ืก ืžืฉืชืžืฉื™ื
09:40
and empowered user base,
261
580135
1694
ืžื™ื•ื“ืข ื•ื‘ืขืœ ื›ื•ื—,
09:41
and I think all of us can agree
262
581829
1100
ื•ืื ื™ ื—ื•ืฉื‘ืช ืฉื›ื•ืœื ื• ื™ื›ื•ืœื™ื ืœื”ืกื›ื™ื
09:42
that that's a pretty ideal way to go forward.
263
582929
2564
ืฉื–ื” ื“ืจืš ืžืื•ื“ ืื™ื“ื™ืืœื™ืช ืœื”ืชืงื“ื.
09:45
Thank you.
264
585493
2184
ืชื•ื“ื” ืœื›ื.
09:47
(Applause)
265
587677
3080
(ืžื—ื™ืื•ืช ื›ืคื™ื™ื)
ืขืœ ืืชืจ ื–ื”

ืืชืจ ื–ื” ื™ืฆื™ื’ ื‘ืคื ื™ื›ื ืกืจื˜ื•ื ื™ YouTube ื”ืžื•ืขื™ืœื™ื ืœืœื™ืžื•ื“ ืื ื’ืœื™ืช. ืชื•ื›ืœื• ืœืจืื•ืช ืฉื™ืขื•ืจื™ ืื ื’ืœื™ืช ื”ืžื•ืขื‘ืจื™ื ืขืœ ื™ื“ื™ ืžื•ืจื™ื ืžื”ืฉื•ืจื” ื”ืจืืฉื•ื ื” ืžืจื—ื‘ื™ ื”ืขื•ืœื. ืœื—ืฅ ืคืขืžื™ื™ื ืขืœ ื”ื›ืชื•ื‘ื™ื•ืช ื‘ืื ื’ืœื™ืช ื”ืžื•ืฆื’ื•ืช ื‘ื›ืœ ื“ืฃ ื•ื™ื“ืื• ื›ื“ื™ ืœื”ืคืขื™ืœ ืืช ื”ืกืจื˜ื•ืŸ ืžืฉื. ื”ื›ืชื•ื‘ื™ื•ืช ื’ื•ืœืœื•ืช ื‘ืกื ื›ืจื•ืŸ ืขื ื”ืคืขืœืช ื”ื•ื•ื™ื“ืื•. ืื ื™ืฉ ืœืš ื”ืขืจื•ืช ืื• ื‘ืงืฉื•ืช, ืื ื ืฆื•ืจ ืื™ืชื ื• ืงืฉืจ ื‘ืืžืฆืขื•ืช ื˜ื•ืคืก ื™ืฆื™ืจืช ืงืฉืจ ื–ื”.

https://forms.gle/WvT1wiN1qDtmnspy7