Or put another way, this time with a more dystopian slant, our social media
platforms are designed for propaganda.
Algorithmic Recommendation: The Cause
of, and Solution to, All of Life’s Problems
Let’s take a step back and get a general view on how this works. (We’ll go into
much more detail in Chapter 3.) Users want content. Whether that’s text,
music, TV, movies, all of that media demand is at its essence a demand for
information. However, information is everywhere. As high as our demand for
these various media is, supply far outstrips it. Both the amount of content and
the ease with which we can access that content ensure that there’s always
plenty of free or inexpensive media we can access.
The problem, of course, is that content producers need to get paid, or they
can’t produce that content. (And while many content producers do it as a side
hustle, our society would surely be impoverished if there were no longer any
filmmakers, songwriters, novelists, or journalists who embarked on their craft
as a full-time, long-term career.) The supply-and-demand equation is so in
favor of the consumer that the average cost of creating media is rapidly
approaching null.17 The amazing thing, though, is that people are still spending
money—a lot of it—on media. But as supply and access continue to increase,
the chance of getting a piece of that pie is rapidly diminishing. So for a content
producer to make money, they need to cut through the noise and be the ones
to get our attention.
15Mike Allen, “Sean Parker unloads on Facebook: ‘God only knows what it’s doing to our
children’s brains’,” Axios, published November 9, 2017, www.axios.com/sean-parker-
unloads-on-facebook-god-only-knows-what-its-doing-to-our-childrens-
brains-1513306792-f855e7b4-4e99-4d60-8d51-2775559c2671.html.
16Ibid.
17Cory Doctorow, Information Doesn’t Want to be Free: Laws for the Internet Age, San
Francisco: McSweeney’s (2014), p. 55ff. See also Paul Mason, Postcapitalism: A Guide to Our
future, New York: Farrar, Straus and Giroux (2015), p. 119.
Data versus Democracy
13
Media consumers want content. Media producers need consumers’ attention.
More than that, media consumers want to cut through all the crap and find
the good stuff—whether that’s entertainment, education, or news. And media
producers need to get their content in front of the right audience—the ones
that might actually care enough to spend their hard-earned money on it.
Consumers want to find the right media. Creators want to find the right
audience. And they’re trying to find each other on the web. Kind of sounds
like a dating site. In some ways, that’s exactly what it is.
At its core, a dating site is a recommendation engine. By taking data about
you and comparing it with data about other people, it recommends people
with whom you might be compatible. The input data may come from an
extended survey or a psychological personality test, or the input data could
simply come from your behavior—hundreds of swipes, left and right. But
ultimately, the process is the same. Take input data, run it through an algorithm,
make recommendations. (And, in the better systems, collect data about how
good the recommendation was, so the algorithm can be improved.)
When it comes to matching consumers with content, or media producers
with audiences, the process is the same. Fine-tune an algorithm that matches
a user with content, maximizing the compatibility so that consumers are
happy with their choice, and producers get paid for their labor. The big
difference with a dating site, though, is that if a dating site is successful, and
you find the love of your life, you never need the dating site again. But if the
media recommendation engine is successful, you’ll not only find a great movie
Friday night, you’ll also come back for another one on Saturday. And remember,
the platforms are also competing for our attention. Netflix, Hulu, Amazon,
Pandora, Spotify, Twitter, Facebook, … the more they get us to come back,
the more money they make, too, regardless of what media we consume.
This seems like a win-win-win situation. Platforms make recommendations;
we find “personalized” entertainment choices that maximize our viewing/
audience for their work, maximizing their income with minimal effort; and
every time we give that track a thumbs up or thumbs down or rate that
movie, the recommendations get better—for us and for the content
producers. And as the recommendations get better, the platforms yield their
own rewards. (When a media platform is also in the content creation business,
it’s a double win for them.)
But the system is not without its flaws. I’ve already discussed how it essentially
primes us for addiction, prompting us to come back unconsciously, out of
habit, rather than deliberately. But there are other problems. In order to
make the best recommendations, platforms need lots of data about both us
14
Chapter 1 | Pay Attention
and the content to feed their algorithms. As platform competition gets fiercer,
and computational resources cheaper, the amount of input data collected
about us increases rapidly.
This data collection often happens without our knowledge. We may agree to
the Terms of Service, but the details implied by those terms are not always
transparent, and they sometimes change after we’re already pretty well baked
into the platform. This data collection also happens when our guard is down,
when we’re the least vigilant about what information we’re providing to whom
and to what ends. This isn’t our bank statement, our mortgage agreement,
our major term papers—the things we are careful and deliberate about. This
is what we read first thing in the morning, before our first cup of coffee,
sometimes before we get out of bed. This is what we listen to as we drive to
work or go for a run. This is what we watch while we have a drink before bed.
We’re talking the music we listen to, the shows we watch, the news we read,
what we say about it, the friends’ pictures of their kids that we “like,” the
news stories that we respond to with a rage emoji, the pictures we spend the
most time looking at, never mind the ads we click (and, in some cases, the
credit card purchases that follow). All of that is logged, and much of it is used,
even traded or sold, by the platforms, in order to serve up the content most
likely to keep us on their platforms the longest and coming back the most
often. This data collection leads to hacks, breaches, leaks, and, in some cases,
targeted ads that know a bit too much about us. (Remember the case a few
years back of the parent who found out their teenage daughter was pregnant
because of the advertisements they received, before their daughter broke the
news?18)
There’s another problem. For a platform to maximize our attention, and their
data collection, they often strive to be one-stop shops. Amazon, Facebook,
and Google (and before them,
Yahoo!) are perhaps the best examples of this.
The same “news feed” that gives us our news also gives us updates from our
extended family, and for some serves as a professional development network.
Then the ads encourage us to join groups where we can connect with people
who share our religious convictions, or simply to buy a new pair of shoes or
an upgraded smartphone. By maximizing our attention through such diverse
content, the platform also divides our attention. They push us deeper into a
state of “continuous partial attention,” as Linda Stone calls it, where no one
thing dominates our thinking. 19 This not only keeps us from thinking slowly
18Kashmir Hill, “How Target Figured Out a Teen Girl Was Pregnant Before Her Father Did,”
Forbes, published February 16, 2012, www.forbes.com/sites/kashmirhill/2012/02/16/
how-target-figured-out-a-teen-girl-was-pregnant-before-her-father-did/.
19Linda Stone, “Beyond Simple Multi-Tasking: Continuous Partial Attention,” Linda Stone
(blog), November 30, 2009, https://lindastone.net/2009/11/30/beyond-simple-
multi-tasking-continuous-partial-attention/.
Data versus Democracy
15
and deeply about any one thing, but it causes us to constantly shift back and
forth between different focuses. The resulting “attentional blink” (as some
cognitive psychologists call it)20 is a time of once again finding our bearings.
And when we’re constantly catching our breath, finding our place on the map,
switching cognitive tasks, we find ourselves in what Stone calls “an artificial
sense of constant crisis.” And that’s not just a psychological problem, that’s a
thinking problem. We can’t slow down, we can’t dive deep, we can’t think
critically. And the more time we spend on these platforms, the more time we
spend in “attentional blink,” in constant cognitive crisis.
When we find ourselves regularly in this psychological state on the platforms
where we find much of our news, we are perfectly primed for propaganda. But
just what is propaganda?
Propaganda Defined
The word propaganda comes from the word propagate—to spread. In its
oldest context, it simply refers to the spreading of a message, whether through
word of mouth or through print media. It’s similar both to publishing (making
public) and to evangelizing (sharing good news), in that sense. But in more
modern times, it’s taken on a more sinister tone. In his classic text Propaganda:
The Formation of Men’s Attitudes, Jacques Ellul writes:
Propaganda is a set of methods employed by an organized group
that wants to bring about the active or passive participation in its
actions of a mass of individuals, psychologically unified through psy-
chological manipulation and incorporated in an organization. 21
I find this definition helpful, but also insufficient for the digital age. The idea of
an organization being the core agent, and expansion of that organization being
the goal, only accounts for a small part of the propaganda activities we see
online. The idea of being a card-carrying member of an organization has
largely been supplanted these days by participation in a movement, with
various degrees of possible participation. This difference in what movement
“membership” entails, as well as the different kinds of messages and media
available to modern citizens, requires some different nuances in how we
define propaganda.
20Howard Rheingold, Net Smart: How to Thrive Online, Cambridge, Mass.: MIT Press (2014),
p. 39.
21Jacques Ellul, Propaganda: The Formation of Men’s Attitudes, New York, Vintage Books
(1965), p. 61.
16
Chapter 1 | Pay Attention
With that in mind, I define propaganda as the use of one or more media to
communicate a message, with the aim of changing someone’s mind or actions
via psychological manipulation, rather than reasoned discourse. Non-propaganda
is not the absence of bias—we’re all biased. Propaganda is the (usually
purposeful) attempt to hide the bias, to present non-facts as facts, to present
facts incompletely or stripped of their essential context, to steer the mind
away from the processes of reason that allow us to read through bias critically
and to discern facts from fiction, truth from lies. Propaganda can involve
disinformation (from the Russian dezinformatsiya)—a purposeful attempt to
deceive or manipulate—or misinformation, an inadvertent spreading of
falsehoods and fallacies. Online, we often see both working in concert—a
purposeful attempt to deceive, shared in such a way that the deceived help to
propagate it, but in earnest. I call this multistage propaganda information
laundering, since the original disinformation is laundered through well-meaning
people, whose activity both spreads the message and obscures the source.
With these definitions, we can see how media addiction, invasive data
collection, and constant “attentional blink” all prime us to be victims of
information operations. The more we spend time on platforms that promote
superficial thinking, the less critically we examine the information we engage
and the sources from which it comes. The more “social” our media
consumption behaviors, the more we let our guard down. The more we shift
cognitive gears, the less capable we are of going deep when we need to. And
the more data is collected about us, the more sophisticated and personally
targeted those operations can be. Harold D. Lasswell writes that propaganda
seeks to subordinate others “while reducing the material cost to power,” 22
and our modern, algorithmically based media platforms provide perhaps the
greatest opportunity in human history to accomplish that cost reduction.
But all hope is not lost. Ellul writes: “Propaganda renders the true exercise of
[democracy] almost impossible” (p. xvi). And as social media feeds the
propaganda machine, many critics are sounding the death knell of modern
democracy. However, the very tools that facilitate the information operations
increasing social polarization and potentially swinging elections are the same
tools that can help us resist. Ellul also writes that because propaganda
dehumanizes and reduces our personal and collective agency, “Propaganda
ceases where simple dialogue begins” (p. 6). And where better to start that
dialogue than on social media?
I understand. It can be easy to dismiss social media, even the web, and pine
for the older forms of human connection (none of which have actually
disappeared, by the way). But as new media scholar Clay Shirky puts it,
22Cited in Jacques Ellul, Propaganda: The Formation of Men’s Attitudes, p. x.
Data versus Democracy
17
“The change we are in the middle of isn’t minor and it isn’t optional, but nor
are its contours set in stone.” He continues:
Our older habits of consumption weren’t virtuous, they were just a
side-effect of living in an environment of impoverished access.
Nostalgia for the accidental scarcity we’ve just emerged from is just
a sideshow; the main event is trying to shape the greatest expansion
of expressive capability the world has ever known.23
The media landscape is a very different place than it was even just ten years
ago. As the constant updates to our social media feeds remind us regularly,
nothing is set in stone in this Brave New World. Not yet. We have the
opportunity to shape it, as both consumers and producers of the content that
keeps those platforms afloat.
But to do that work of resistance, of redrawing the blueprints of media and
society, first we need to know how it works. In the next two chapters, we’ll
dive deeply into how we work when we engage information, and then into
how the systems work. That knowledge, plus the examples—both hopeful and
tragic—explored in the latter half of this book, will give us the foothold we
need as we seek to solve the propaganda problem.
But for now, I’ll leave you with some of the best new media advice I’ve ever
heard.
Throw some sand into the machinery that automatizes your
attention.
—Howard Rheingold24
Summary
In this chapter we’ve learned that Western capitalism has moved from a
commodity-based economy to an attention-based economy. The supply-and-
demand equation that increasingly governs the way we interact with
information deals with the limited supply and increasing demand of human
attention, rather than information, goods, or services.
Algorithmic recommendation engines and social media feeds have been
created to help users find the most relevant content and to help media
23Clay Shirky, “Why Abundance Is Good: A Reply to Nick Carr,” Encyclopaedia Britannica
Blog, July 17, 2008, http://blogs.britannica.com/2008/07/why-abundance-is-good-
a-reply-to-nick-carr/.
24 Net Smart, p. 50.
18
Chapter 1 | Pay Attention
producers find the most appropriate audiences. But the ways in which media
producers compete for our attention, the amount of personal data mined to
make the algorithms work, and the natural way our cognitive systems function
all combine to make the modern media landscape ripe for propaganda.
However, if we understand the economy, biology, and technology, we can
begin to counter the negative effects and even use the same tools to undo the