TED2019
Claire Wardle: How you can help transform the internet into a place of trust
Filmed:
Readability: 4.8
1,319,606 views
How can we stop the spread of misleading, sometimes dangerous content while maintaining an internet with freedom of expression at its core? Misinformation expert Claire Wardle explores the new challenges of our polluted online environment and maps out a plan to transform the internet into a place of trust -- with the help everyday users. "Together, let's rebuild our information commons," she says.
Claire Wardle - Misinformation expert
Claire Wardle is an expert on user-generated content and verification working to help improve the quality of information online. Full bio
Claire Wardle is an expert on user-generated content and verification working to help improve the quality of information online. Full bio
Double-click the English transcript below to play the video.
00:13
No matter who you are or where you live,
0
1548
2516
00:16
I'm guessing that you have
at least one relative
at least one relative
1
4088
2356
00:18
that likes to forward those emails.
2
6468
2334
00:21
You know the ones I'm talking about --
3
9206
1900
00:23
the ones with dubious claims
or conspiracy videos.
or conspiracy videos.
4
11130
2854
00:26
And you've probably
already muted them on Facebook
already muted them on Facebook
5
14315
2668
00:29
for sharing social posts like this one.
6
17007
2348
00:31
It's an image of a banana
7
19379
1404
00:32
with a strange red cross
running through the center.
running through the center.
8
20807
2667
00:35
And the text around it is warning people
9
23498
2139
00:37
not to eat fruits that look like this,
10
25661
2158
00:39
suggesting they've been
injected with blood
injected with blood
11
27843
2028
00:41
contaminated with the HIV virus.
12
29895
2130
00:44
And the social share message
above it simply says,
above it simply says,
13
32049
2603
00:46
"Please forward to save lives."
14
34676
2181
00:49
Now, fact-checkers have been debunking
this one for years,
this one for years,
15
37672
3294
00:52
but it's one of those rumors
that just won't die.
that just won't die.
16
40990
2809
00:55
A zombie rumor.
17
43823
1271
00:57
And, of course, it's entirely false.
18
45513
2093
01:00
It might be tempting to laugh
at an example like this, to say,
at an example like this, to say,
19
48180
2959
01:03
"Well, who would believe this, anyway?"
20
51163
1884
01:05
But the reason it's a zombie rumor
21
53419
1626
01:07
is because it taps into people's
deepest fears about their own safety
deepest fears about their own safety
22
55069
3889
01:10
and that of the people they love.
23
58982
2175
01:13
And if you spend as enough time
as I have looking at misinformation,
as I have looking at misinformation,
24
61783
3273
01:17
you know that this is just
one example of many
one example of many
25
65080
2420
01:19
that taps into people's deepest
fears and vulnerabilities.
fears and vulnerabilities.
26
67524
3047
01:23
Every day, across the world,
we see scores of new memes on Instagram
we see scores of new memes on Instagram
27
71214
4369
01:27
encouraging parents
not to vaccinate their children.
not to vaccinate their children.
28
75607
3039
01:30
We see new videos on YouTube
explaining that climate change is a hoax.
explaining that climate change is a hoax.
29
78670
4532
01:35
And across all platforms, we see
endless posts designed to demonize others
endless posts designed to demonize others
30
83226
4302
01:39
on the basis of their race,
religion or sexuality.
religion or sexuality.
31
87552
3501
01:44
Welcome to one of the central
challenges of our time.
challenges of our time.
32
92314
3030
01:47
How can we maintain an internet
with freedom of expression at the core,
with freedom of expression at the core,
33
95647
4025
01:51
while also ensuring that the content
that's being disseminated
that's being disseminated
34
99696
3303
01:55
doesn't cause irreparable harms
to our democracies, our communities
to our democracies, our communities
35
103023
3886
01:58
and to our physical and mental well-being?
36
106933
2238
02:01
Because we live in the information age,
37
109998
2087
02:04
yet the central currency
upon which we all depend -- information --
upon which we all depend -- information --
38
112109
3547
02:07
is no longer deemed entirely trustworthy
39
115680
2357
02:10
and, at times, can appear
downright dangerous.
downright dangerous.
40
118061
2328
02:12
This is thanks in part to the runaway
growth of social sharing platforms
growth of social sharing platforms
41
120811
3937
02:16
that allow us to scroll through,
42
124772
1642
02:18
where lies and facts sit side by side,
43
126438
2222
02:20
but with none of the traditional
signals of trustworthiness.
signals of trustworthiness.
44
128684
3071
02:24
And goodness -- our language around this
is horribly muddled.
is horribly muddled.
45
132268
3619
02:27
People are still obsessed
with the phrase "fake news,"
with the phrase "fake news,"
46
135911
3103
02:31
despite the fact that
it's extraordinarily unhelpful
it's extraordinarily unhelpful
47
139038
2531
02:33
and used to describe a number of things
that are actually very different:
that are actually very different:
48
141593
3460
02:37
lies, rumors, hoaxes,
conspiracies, propaganda.
conspiracies, propaganda.
49
145077
3386
02:40
And I really wish
we could stop using a phrase
we could stop using a phrase
50
148911
2912
02:43
that's been co-opted by politicians
right around the world,
right around the world,
51
151847
2862
02:46
from the left and the right,
52
154733
1471
02:48
used as a weapon to attack
a free and independent press.
a free and independent press.
53
156228
3222
02:52
(Applause)
54
160307
4702
02:57
Because we need our professional
news media now more than ever.
news media now more than ever.
55
165033
3462
03:00
And besides, most of this content
doesn't even masquerade as news.
doesn't even masquerade as news.
56
168882
3373
03:04
It's memes, videos, social posts.
57
172279
2642
03:06
And most of it is not fake;
it's misleading.
it's misleading.
58
174945
3453
03:10
We tend to fixate on what's true or false.
59
178422
3015
03:13
But the biggest concern is actually
the weaponization of context.
the weaponization of context.
60
181461
4032
03:18
Because the most effective disinformation
61
186855
1968
03:20
has always been that
which has a kernel of truth to it.
which has a kernel of truth to it.
62
188847
3048
03:23
Let's take this example
from London, from March 2017,
from London, from March 2017,
63
191919
3476
03:27
a tweet that circulated widely
64
195419
1540
03:28
in the aftermath of a terrorist incident
on Westminster Bridge.
on Westminster Bridge.
65
196983
3587
03:32
This is a genuine image, not fake.
66
200594
2428
03:35
The woman who appears in the photograph
was interviewed afterwards,
was interviewed afterwards,
67
203046
3169
03:38
and she explained that
she was utterly traumatized.
she was utterly traumatized.
68
206239
2409
03:40
She was on the phone to a loved one,
69
208672
1738
03:42
and she wasn't looking
at the victim out of respect.
at the victim out of respect.
70
210434
2618
03:45
But it still was circulated widely
with this Islamophobic framing,
with this Islamophobic framing,
71
213076
3960
03:49
with multiple hashtags,
including: #BanIslam.
including: #BanIslam.
72
217060
3046
03:52
Now, if you worked at Twitter,
what would you do?
what would you do?
73
220425
2398
03:54
Would you take that down,
or would you leave it up?
or would you leave it up?
74
222847
2562
03:58
My gut reaction, my emotional reaction,
is to take this down.
is to take this down.
75
226553
3429
04:02
I hate the framing of this image.
76
230006
2142
04:04
But freedom of expression
is a human right,
is a human right,
77
232585
2388
04:06
and if we start taking down speech
that makes us feel uncomfortable,
that makes us feel uncomfortable,
78
234997
3225
04:10
we're in trouble.
79
238246
1230
04:11
And this might look like a clear-cut case,
80
239500
2294
04:13
but, actually, most speech isn't.
81
241818
1698
04:15
These lines are incredibly
difficult to draw.
difficult to draw.
82
243540
2436
04:18
What's a well-meaning
decision by one person
decision by one person
83
246000
2281
04:20
is outright censorship to the next.
84
248305
2077
04:22
What we now know is that
this account, Texas Lone Star,
this account, Texas Lone Star,
85
250759
2929
04:25
was part of a wider Russian
disinformation campaign,
disinformation campaign,
86
253712
3230
04:28
one that has since been taken down.
87
256966
2151
04:31
Would that change your view?
88
259141
1563
04:33
It would mine,
89
261322
1159
04:34
because now it's a case
of a coordinated campaign
of a coordinated campaign
90
262505
2301
04:36
to sow discord.
91
264830
1215
04:38
And for those of you who'd like to think
92
266069
1961
04:40
that artificial intelligence
will solve all of our problems,
will solve all of our problems,
93
268054
2831
04:42
I think we can agree
that we're a long way away
that we're a long way away
94
270909
2225
04:45
from AI that's able to make sense
of posts like this.
of posts like this.
95
273158
2587
04:48
So I'd like to explain
three interlocking issues
three interlocking issues
96
276856
2507
04:51
that make this so complex
97
279387
2373
04:53
and then think about some ways
we can consider these challenges.
we can consider these challenges.
98
281784
3122
04:57
First, we just don't have
a rational relationship to information,
a rational relationship to information,
99
285348
3890
05:01
we have an emotional one.
100
289262
1468
05:02
It's just not true that more facts
will make everything OK,
will make everything OK,
101
290754
3794
05:06
because the algorithms that determine
what content we see,
what content we see,
102
294572
3100
05:09
well, they're designed to reward
our emotional responses.
our emotional responses.
103
297696
3127
05:12
And when we're fearful,
104
300847
1381
05:14
oversimplified narratives,
conspiratorial explanations
conspiratorial explanations
105
302252
3174
05:17
and language that demonizes others
is far more effective.
is far more effective.
106
305450
3418
05:21
And besides, many of these companies,
107
309538
1874
05:23
their business model
is attached to attention,
is attached to attention,
108
311436
2546
05:26
which means these algorithms
will always be skewed towards emotion.
will always be skewed towards emotion.
109
314006
3690
05:30
Second, most of the speech
I'm talking about here is legal.
I'm talking about here is legal.
110
318371
4298
05:35
It would be a different matter
111
323081
1446
05:36
if I was talking about
child sexual abuse imagery
child sexual abuse imagery
112
324551
2341
05:38
or content that incites violence.
113
326916
1927
05:40
It can be perfectly legal
to post an outright lie.
to post an outright lie.
114
328867
3270
05:45
But people keep talking about taking down
"problematic" or "harmful" content,
"problematic" or "harmful" content,
115
333130
4034
05:49
but with no clear definition
of what they mean by that,
of what they mean by that,
116
337188
2609
05:51
including Mark Zuckerberg,
117
339821
1264
05:53
who recently called for global
regulation to moderate speech.
regulation to moderate speech.
118
341109
3412
05:56
And my concern is that
we're seeing governments
we're seeing governments
119
344870
2215
05:59
right around the world
120
347109
1292
06:00
rolling out hasty policy decisions
121
348425
2676
06:03
that might actually trigger
much more serious consequences
much more serious consequences
122
351125
2746
06:05
when it comes to our speech.
123
353895
1714
06:08
And even if we could decide
which speech to take up or take down,
which speech to take up or take down,
124
356006
3706
06:11
we've never had so much speech.
125
359736
2174
06:13
Every second, millions
of pieces of content
of pieces of content
126
361934
2131
06:16
are uploaded by people
right around the world
right around the world
127
364089
2107
06:18
in different languages,
128
366220
1168
06:19
drawing on thousands
of different cultural contexts.
of different cultural contexts.
129
367412
2768
06:22
We've simply never had
effective mechanisms
effective mechanisms
130
370204
2532
06:24
to moderate speech at this scale,
131
372760
1738
06:26
whether powered by humans
or by technology.
or by technology.
132
374522
2801
06:30
And third, these companies --
Google, Twitter, Facebook, WhatsApp --
Google, Twitter, Facebook, WhatsApp --
133
378284
3944
06:34
they're part of a wider
information ecosystem.
information ecosystem.
134
382252
2841
06:37
We like to lay all the blame
at their feet, but the truth is,
at their feet, but the truth is,
135
385117
3352
06:40
the mass media and elected officials
can also play an equal role
can also play an equal role
136
388493
3830
06:44
in amplifying rumors and conspiracies
when they want to.
when they want to.
137
392347
2913
06:47
As can we, when we mindlessly forward
divisive or misleading content
divisive or misleading content
138
395800
4944
06:52
without trying.
139
400768
1285
06:54
We're adding to the pollution.
140
402077
1800
06:57
I know we're all looking for an easy fix.
141
405236
2618
06:59
But there just isn't one.
142
407878
1667
07:01
Any solution will have to be rolled out
at a massive scale, internet scale,
at a massive scale, internet scale,
143
409950
4445
07:06
and yes, the platforms,
they're used to operating at that level.
they're used to operating at that level.
144
414419
3261
07:09
But can and should we allow them
to fix these problems?
to fix these problems?
145
417704
3472
07:13
They're certainly trying.
146
421668
1232
07:14
But most of us would agree that, actually,
we don't want global corporations
we don't want global corporations
147
422924
4086
07:19
to be the guardians of truth
and fairness online.
and fairness online.
148
427034
2332
07:21
And I also think the platforms
would agree with that.
would agree with that.
149
429390
2537
07:24
And at the moment,
they're marking their own homework.
they're marking their own homework.
150
432257
2881
07:27
They like to tell us
151
435162
1198
07:28
that the interventions
they're rolling out are working,
they're rolling out are working,
152
436384
2579
07:30
but because they write
their own transparency reports,
their own transparency reports,
153
438987
2540
07:33
there's no way for us to independently
verify what's actually happening.
verify what's actually happening.
154
441551
3618
07:38
(Applause)
155
446431
3342
07:41
And let's also be clear
that most of the changes we see
that most of the changes we see
156
449797
2952
07:44
only happen after journalists
undertake an investigation
undertake an investigation
157
452773
2994
07:47
and find evidence of bias
158
455791
1611
07:49
or content that breaks
their community guidelines.
their community guidelines.
159
457426
2829
07:52
So yes, these companies have to play
a really important role in this process,
a really important role in this process,
160
460815
4595
07:57
but they can't control it.
161
465434
1560
07:59
So what about governments?
162
467855
1518
08:01
Many people believe
that global regulation is our last hope
that global regulation is our last hope
163
469863
3096
08:04
in terms of cleaning up
our information ecosystem.
our information ecosystem.
164
472983
2880
08:07
But what I see are lawmakers
who are struggling to keep up to date
who are struggling to keep up to date
165
475887
3166
08:11
with the rapid changes in technology.
166
479077
2341
08:13
And worse, they're working in the dark,
167
481442
1904
08:15
because they don't have access to data
168
483370
1821
08:17
to understand what's happening
on these platforms.
on these platforms.
169
485215
2650
08:20
And anyway, which governments
would we trust to do this?
would we trust to do this?
170
488260
3071
08:23
We need a global response,
not a national one.
not a national one.
171
491355
2770
08:27
So the missing link is us.
172
495419
2277
08:29
It's those people who use
these technologies every day.
these technologies every day.
173
497720
3123
08:33
Can we design a new infrastructure
to support quality information?
to support quality information?
174
501260
4591
08:38
Well, I believe we can,
175
506371
1230
08:39
and I've got a few ideas about
what we might be able to actually do.
what we might be able to actually do.
176
507625
3357
08:43
So firstly, if we're serious
about bringing the public into this,
about bringing the public into this,
177
511006
3103
08:46
can we take some inspiration
from Wikipedia?
from Wikipedia?
178
514133
2381
08:48
They've shown us what's possible.
179
516538
1824
08:50
Yes, it's not perfect,
180
518386
1151
08:51
but they've demonstrated
that with the right structures,
that with the right structures,
181
519561
2634
08:54
with a global outlook
and lots and lots of transparency,
and lots and lots of transparency,
182
522219
2635
08:56
you can build something
that will earn the trust of most people.
that will earn the trust of most people.
183
524878
3096
08:59
Because we have to find a way
to tap into the collective wisdom
to tap into the collective wisdom
184
527998
3162
09:03
and experience of all users.
185
531184
2309
09:05
This is particularly the case
for women, people of color
for women, people of color
186
533517
2646
09:08
and underrepresented groups.
187
536187
1346
09:09
Because guess what?
188
537557
1166
09:10
They are experts when it comes
to hate and disinformation,
to hate and disinformation,
189
538747
2735
09:13
because they have been the targets
of these campaigns for so long.
of these campaigns for so long.
190
541506
3516
09:17
And over the years,
they've been raising flags,
they've been raising flags,
191
545046
2350
09:19
and they haven't been listened to.
192
547420
1665
09:21
This has got to change.
193
549109
1280
09:22
So could we build a Wikipedia for trust?
194
550807
4326
09:27
Could we find a way that users
can actually provide insights?
can actually provide insights?
195
555157
4189
09:31
They could offer insights around
difficult content-moderation decisions.
difficult content-moderation decisions.
196
559370
3697
09:35
They could provide feedback
197
563091
1463
09:36
when platforms decide
they want to roll out new changes.
they want to roll out new changes.
198
564578
3041
09:40
Second, people's experiences
with the information is personalized.
with the information is personalized.
199
568241
4162
09:44
My Facebook news feed
is very different to yours.
is very different to yours.
200
572427
2643
09:47
Your YouTube recommendations
are very different to mine.
are very different to mine.
201
575094
2745
09:49
That makes it impossible for us
to actually examine
to actually examine
202
577863
2492
09:52
what information people are seeing.
203
580379
2023
09:54
So could we imagine
204
582815
1389
09:56
developing some kind of centralized
open repository for anonymized data,
open repository for anonymized data,
205
584228
4778
10:01
with privacy and ethical
concerns built in?
concerns built in?
206
589030
2864
10:04
Because imagine what we would learn
207
592220
1778
10:06
if we built out a global network
of concerned citizens
of concerned citizens
208
594022
3261
10:09
who wanted to donate
their social data to science.
their social data to science.
209
597307
3294
10:13
Because we actually know very little
210
601141
1722
10:14
about the long-term consequences
of hate and disinformation
of hate and disinformation
211
602887
2881
10:17
on people's attitudes and behaviors.
212
605792
1975
10:20
And what we do know,
213
608236
1167
10:21
most of that has been
carried out in the US,
carried out in the US,
214
609427
2193
10:23
despite the fact that
this is a global problem.
this is a global problem.
215
611644
2381
10:26
We need to work on that, too.
216
614049
1635
10:28
And third,
217
616192
1150
10:29
can we find a way to connect the dots?
218
617366
2310
10:31
No one sector, let alone nonprofit,
start-up or government,
start-up or government,
219
619700
3438
10:35
is going to solve this.
220
623162
1422
10:36
But there are very smart people
right around the world
right around the world
221
624608
2564
10:39
working on these challenges,
222
627196
1381
10:40
from newsrooms, civil society,
academia, activist groups.
academia, activist groups.
223
628601
3576
10:44
And you can see some of them here.
224
632201
1898
10:46
Some are building out indicators
of content credibility.
of content credibility.
225
634123
2927
10:49
Others are fact-checking,
226
637074
1246
10:50
so that false claims, videos and images
can be down-ranked by the platforms.
can be down-ranked by the platforms.
227
638344
3591
10:53
A nonprofit I helped
to found, First Draft,
to found, First Draft,
228
641959
2213
10:56
is working with normally competitive
newsrooms around the world
newsrooms around the world
229
644196
2968
10:59
to help them build out investigative,
collaborative programs.
collaborative programs.
230
647188
3503
11:03
And Danny Hillis, a software architect,
231
651231
2309
11:05
is designing a new system
called The Underlay,
called The Underlay,
232
653564
2381
11:07
which will be a record
of all public statements of fact
of all public statements of fact
233
655969
2775
11:10
connected to their sources,
234
658768
1329
11:12
so that people and algorithms
can better judge what is credible.
can better judge what is credible.
235
660121
3654
11:16
And educators around the world
are testing different techniques
are testing different techniques
236
664800
3356
11:20
for finding ways to make people
critical of the content they consume.
critical of the content they consume.
237
668180
3484
11:24
All of these efforts are wonderful,
but they're working in silos,
but they're working in silos,
238
672633
3141
11:27
and many of them are woefully underfunded.
239
675798
2680
11:30
There are also hundreds
of very smart people
of very smart people
240
678502
2053
11:32
working inside these companies,
241
680579
1652
11:34
but again, these efforts
can feel disjointed,
can feel disjointed,
242
682255
2325
11:36
because they're actually developing
different solutions to the same problems.
different solutions to the same problems.
243
684604
3937
11:41
How can we find a way
to bring people together
to bring people together
244
689205
2269
11:43
in one physical location
for days or weeks at a time,
for days or weeks at a time,
245
691498
3278
11:46
so they can actually tackle
these problems together
these problems together
246
694800
2396
11:49
but from their different perspectives?
247
697220
1818
11:51
So can we do this?
248
699062
1340
11:52
Can we build out a coordinated,
ambitious response,
ambitious response,
249
700426
3239
11:55
one that matches the scale
and the complexity of the problem?
and the complexity of the problem?
250
703689
3709
11:59
I really think we can.
251
707819
1373
12:01
Together, let's rebuild
our information commons.
our information commons.
252
709216
2959
12:04
Thank you.
253
712819
1190
12:06
(Applause)
254
714033
3728
ABOUT THE SPEAKER
Claire Wardle - Misinformation expertClaire Wardle is an expert on user-generated content and verification working to help improve the quality of information online.
Why you should listen
Claire Wardle is the executive director of First Draft, a nonprofit dedicated to educating journalists about reporting in an age of information disorder. Previously, she was a research fellow at the Shorenstein Center on Media, Politics and Public Policy at the Harvard Kennedy School; the research director at the Tow Center for Digital Journalism at Columbia Journalism School; the head of social media for the UN Refugee Agency; and director of news services for Storyful. Wardle holds a PhD in communication and an MA in political science from the University of Pennsylvania.
Claire Wardle | Speaker | TED.com