by Kris Shaffer
Americans who held views that, no matter how real and personal, happened to
align with the Kremlin’s interests. By reinforcing existing social biases and
personal interests early on, the IRA were able to “hack” the model so that
users most likely to engage Kremlin-friendly messages saw more of them,
even if the sources were legitimate.
This is a situation where the contrast between disinformation as manipulative
behavior and disinformation as deceptive messages comes into sharp relief. By
researching and weaponizing existing social biases, the IRA was able to wield
those earnestly held beliefs in their efforts to manipulate unwitting Americans
into taking actions they may or may not have otherwise taken. And that was
true regardless of whether or not the posts and memes were factual.
No amount of individual fact-checking will solve this problem. This is first and
foremost a behavioral issue that can only be seen and solved with the analysis
of data on the large scale: cross-community, cross-platform, over long
stretches of time. But if platforms, governments, and/or third-party researchers
conduct that large-scale analysis, and accounts are suspended for this
coordinated, inorganic, manipulative behavior, then those more sophisticated
state actors lose many of their best tools for manipulating communities into
unwittingly doing their bidding. And in the process, we all regain the integrity
of our information space.
That integrity doesn’t come easy, especially when we are facing influence
operations from sophisticated actors like a military or state-sponsored firm.
As Russia’s operations in particular demonstrate, disinformation is international
and cross-platform and weaponizes fact and fiction alike. As a result, no one
entity has all of the necessary data or expertise to solve the problem.
Platforms don’t have each other’s data, nor the data held by intelligence or law
enforcement agencies. And they do not have the cultural or linguistic experts
that the intelligence agencies, universities, and nonprofit think-tanks do.
Governments don’t have the platforms’ data—and can only collect data under
certain legal circumstances—nor do they have many employees with significant
engineering experience. And while third-party researchers have, at times, a
healthier mix of experts in different areas, they also have less direct access to
platform data. Just like the U.S. government realized after the September 11
attacks, information sharing and the collaborative utilization of a variety of
areas of expertise will be key to addressing the most sophisticated
disinformation campaigns. And given the legal constraints Western
governments rightly face when it comes to surveilling their own citizens,
114
Chapter 7 | Conclusion
and the competitive relationship the tech companies have with each other, it
is likely that carefully structured public-private partnerships will be required
to address the disinformation defense needs that societies around the globe
are increasingly encountering.
But What Can I Do?
It’s clear that disinformation is a large-scale problem that requires large-scale
solutions. Platforms, governments, and researchers must all do their part to
address this rapidly evolving problem collaboratively. But is there anything
that individuals can do?
We’ve already addressed a few ways that individuals can alter the inputs that
big data algorithms use to determine the content that we see on our favorite
platforms. Mindfulness about what we consume, what we engage, what we
share, and what we create is absolutely essential, both to fight “fake news”
and to fight the amplification of social bias and polarizing messaging.
Similarly, we can limit the data about us that individual algorithms have access
to. Consuming some content offline, registering for different services with
different email addresses, using different browsers for different services, even
paying cash and avoiding buyer rewards programs can keep our personal data
in different, smaller piles rather than one big one that many platforms can
access. The result may be less relevant ads (is that really a problem?!), but also
less effective hyper-targeting with messages designed to manipulate.
But ultimately, some networked activism and collective political action will
be necessary. Elected leaders, regulators, and platform executives need to
be involved in solving the propaganda problem, and to do that they need to
be motivated—electorally, legally, financially. But for that to work, we don’t
just need governments and corporations working together, all of us need to
work together.
Democracy depends on the free flow of information, both to inform and to
afford all of us the chance to deliberate and persuade each other. If we can’t
trust the integrity of the information we consume, we can’t trust the
democratic process.
I opened Chapter 6 with a quote from Zeynep Tufekci that bears repeating.
“Technologies alter our ability to preserve and circulate ideas and stories, the
ways in which we connect and converse, the people with whom we can
interact, the things that we can see, and the structures of power that oversee
the means of contact. ”3 Technologies alter the structures of power. That’s not 3Zeynep Tufekci, Twitter and Tear Gas: The Power and Fragility of Networked Protest (New
Haven: Yale University Press, 2017), 5.
Data versus Democracy
115
necessarily a bad thing (remember Ferguson), nor is it necessarily a good thing
(remember GamerGate), but it’s absolutely not a neutral thing. It’s up to us to
invent, implement, and rein in technologies in ways that bring our vision of
what the world could be to fruition.
Too often we are passive, going with the flow, letting the affordances of the
technology—and its manipulation by more motivated, and often less ethical,
actors—determine the future we are building. If we want to make a positive
mark on the universe, we need mindfulness, deliberateness, and collaboration,
as we step into the next chapter of our history. The problem likely won’t go
away anytime soon. But that shouldn’t stop us from working to minimize its
effects and harness the power of new technologies for good.
I
Index
A
4chan, 57, 58, 60–64
Abundance, 4–6
8chan, 57–60
Activists, 48, 49, 51–53, 55, 58, 61, 64
Clickbait, 19
Adaptations, 4
Clinton, Hillary, 67–69, 75, 77, 80, 81, 84, 86
Advertising, 20, 27–29
Cognitive system
clickbait, 19
Advertising, targeted, 110
human brain, 21
Algorithmic neutrality, 34
memory, 21
Algorithmic recommendation, 12–15
processor, 21
Algorithmic timeline, 54
RAM, 20
STM and LTM, 22
Amnesty International, 54
triggers, attention, 22, 23
Application programming interface (API), 96
unconscious memory
Arab Spring, 9
3, 94, 101
Apple, 27
characters, 25
Attentional blink, 109
emotions, 23, 24
Attention economy, 10–12
familiar and unfamiliar things, 25
iPod, 27
B
learn patterns, 26
Big data algorithms, 114
life-or-death evolutionary
history, 24
Black Lives Matter movement, 47–49, 53, 61
marketers, 28
Botnet, 85
perceptual fluency, 24
Bots, 48, 58, 59, 95–97
physical limitations, 24
Brazil, 95–97
pleasurable panting, 23
psychological principle, 27
Breitbart, 60, 62
social media and big
Bronco Bots, 104
data analytics, 28
web technology, 29
C
Collaborative filtering, 36–40, 43, 110
Capitalism, 7, 9
Comey, James, 68, 82
Carbohydrates, 4, 5
Confirmation bias, 109
© Kris Shaffer 2019
K. Shaffer, Data versus Democracy,
https://doi.org/10.1007/978-1-4842-4540-8
118
Index
Content recommendation systems
Feedback loop, 38, 39, 41
bias amplifier
Feudalism, 7, 9
collaborative filtering, 39, 40
feedback loop, 38–40
Food scarcity, 6
polarization, 41
online advertising, 32
G
algorithmic neutrality, 34
Gamelan, 26
collaborative filtering, 43
GamerGate, 55, 59
Google’s search engine, 32–34
information consumption, 43
Genetic code, 5
machine learning model, 34
Global watchdog agencies, 99
optimal engagement, 42
Google, 32–35, 42
promote engagement, 42
stream works, 35–38
H
Crap detector, 6
Homo sapiens, 6
Crash override network, 59
Human rights, 101
Crimea, 70, 71
Human Rights Watch group, 101
Currency, 11, 12
I, J, K
D
Ice Bucket Challenge, 50
Democratic National Committee
Industrial revolution, 6
(DNC), 69, 76, 80–82, 86
Information laundering, 16
Depression Quest, 56, 58
Information warfare, 67, 69, 70, 72,
Digital community center, 52
74, 82, 83
Digital revolution
Internet Research Agency (IRA), 75, 76,
automobile, 92
82–85, 87
Facebook, 94
Project Lakhta
frictionless design, 94
botnet, 85
industrial technology, 92
IRA operations, 87
social structures, 94
organization, 84
Duterte, Rodrigo, 97
L
E
Latin American politics
Economy, 6–9, 12
account profiles, 104
Egypt, 93–95
botnets, 104
Election, 2016 U.S. presidential, 67, 76
network of bots or fake accounts, 104
peace deal, 105
Emotional preference, 4
peer-to-peer messaging, 105, 106
Evolutionary psychology, 4
threat, 103
violence, 104
F
LGBTQ community, 57, 60
Facebook, 47, 50, 54, 63, 96
Long-range acoustic
Fake news, 5
devices (LRADs), 47, 50, 51
Fancy Bear (APT28), 75
Long-term memory (LTM), 21
Index
119
M
Project Lakhta, 82–87
Mainstream media vs. social media, 51
Propaganda, defined, 15–17
Market-based capitalism, 7
Propaganda problem
big data algorithms, 114
Market-driven economy, 7
collaborative filtering, 110
McKesson, DeRay, 48, 49, 52, 53, 64
community mirage, 112
Media, 6, 9–13, 15–18
confirmation bias, 109
Media strategy, 98
disinformation, 110, 113, 114
individual algorithms, 114
Mere exposure effect, 24
media mirage, 112
Microtargeting, 29
pluralistic ignorance, 110
Misinformation, 16
priming, 109
public-private partnerships, 114
Models, 32, 34–38
social stereotypes, 111
Moore’s law, 8
vicious cycle, 111
Myanmar, 95, 99, 102
Psychological warfare, 101, 102, 108
automation, 102
digital technology, 100
Q
ethnic cleansing, 102
Facebook, 101
Quinn, Zoë, 56–59, 64
internet, 100, 101
local rumor mills, 101
R
mass media technology, 100
Random access memory (RAM), 20
mobile phone, 101
violence, 99
Recommendation engine, 13, 17
vulnerability, 102
Reddit, 57, 60, 63, 64
Responsibility, moral/ethical, 64
N, O
Rohingya, 99–102
Natural selection, 5
Rousseff, Dilma, 95, 96
Negative emotions, 4
Russia, 69–75, 77, 80, 82
Russian disinformation operations
P
active measures in Baltic, 72–74
Peer-to-peer disinformation, 107
Crimea, 70, 71
election, 2016 U.S. presidential, 67
Perceptual fluency, 24
information warfare, 67, 69, 70
Philippines, 97, 98
Russian [Twitter] bots, 69
Pizzagate conspiracy theory, 61
Ukrainian separatists, 70–72
Pleistocene period, 4
Russian military intelligence (GRU), 75
APT28
Pluralistic ignorance, 93, 110
DNC, 69, 86
Polarization, 35, 41, 44
hacks of Democratic Party, 76, 77
Police, 49–52, 54, 59–61
Political campaigns, 95
S
Priming, 23
Said, Khaled, 93
Product placement, 27
Scarcity, 17
Professional development, 92
Short-term memory (STM), 22
120
Index
Social-justice warrior (SJWs), 55
U
Social media, 91, 95
Ukraine, 69–72
Social platforms, impact, 64
U.S. Senate Select Committee on Intelligence
Sockpuppets, 48, 58, 63
(SSCI), 82, 83
Stream, 6
Subreddits, 57, 58, 63
V
Supply and demand, 5–10, 12
Verificado 2018, 106, 108
Sweden, 72–74
Vine, 47, 49, 53
VKontakte, 71, 85
T
W, X
Tahrir Square, 93
Weak ties, 110
Trolls, 47, 56–61
WhatsApp, 105–107
Trump, Donald, 68, 69, 75, 80, 81, 83, 86, 87
Tufekci, Zeynep, 92, 93
Y, Z
Twitter, 49
–51, 96, 104
Yiannopoulos, Milo, 60, 62
Document Outline
Contents
About the Author
Acknowledgments
Introduction: From Scarcity to Abundance
Part I: The Propaganda Problem Chapter 1: Pay Attention How Taste Is Made
Supply and Demand: Why an Information Economy Is No Longer Sustainable
If You Don’t Pay for the Product, You Are the Product: Attention as Commodity, Engagement as Currency
Algorithmic Recommendation: The Cause of, and Solution to, All of Life’s Problems
Propaganda Defined
Summary
Chapter 2: Cog in the System Clickbait: You Won’t Believe What Happens Next! Mapping the Cognitive System
The Limits of Conscious Attention The Triggers of Attention
Familiarity Breeds Believability: The Role of Unconscious Memory
Summary
Chapter 3: Swimming Upstream What’s New?
How the Stream Works
Bias Amplifier
Letting Your Guard Down
Summary
Part II: Case Studies Chapter 4: Domestic Disturbance Crowd Control: How the Tweets of Ferguson Steered Mainstream Media and Public Awareness August 9, 2014: What Happened in Ferguson
A Movement Emerges as “Leaderless” Activists Organize on Twitter
Who Decides What Stories Get Told?
You Can’t Just Quit the Internet: How GamerGate Turned Social Media into a “Real-life” Weapon Zoë Quinn and the Blog Post from Hell
Antisocial Media: When Domestic Psychological Abuse Tactics Scale Up
Unprepared: How Platforms, Police, and the Courts (Failed to) Respond
The Emergence of the Alt-Right
The Mob Rules -or- Who Decides What Stories Get Told? [redux]
Summary
Chapter 5: Democracy Hacked, Part 1 What Happened?
Meet the New War, Same as the (C)old War
Ukrainian “Separatists”
Active Measures in the Baltic
Fancy Bear and the Great Meme War of 2016 Fancy Bear Crashes the Democratic Party
How Fancy Bear Got In
Project Lakhta
What Now?
Summary
Chapter 6: Democracy Hacked, Part 2 A Digital Revolution
Bots in Brazil
“Weaponizing” Facebook in the Philippines
Consolidating Power in Myanmar
Success in the Latin American Elections of 2018
Summary
Chapter 7: Conclusion The Propaganda Problem
But What Can I Do?
Index