neon sign, zero like count
PHOTO: Prateek Katyal

When I was a kid, I loved and hated scary movies. I didn’t eat mushrooms for years after seeing a B-movie where eating them turned you into a Toadstool person. Invasion of the Body Snatchers, with the main character named Myles? Even scarier. Against this backdrop, I write this review of "The Hype Machine: How Social Media Disrupts Our Elections, Our Economy, and Our Health — and How We Must Adapt" by Sinan Aral, which publishes on Sept. 15. Honestly, the scenario Aral describes should scare any technology leader, user of social media technology, or citizen within a democracy.

To be fair, any technology can be used for good or bad. As a professor in a graduate school class on futures research said, "We have a choice to how we respond to technology." You have a choice after you read this piece regarding whether you want to maintain the social contract that Rousseau suggested several hundred years ago or not. Because in the end, it is our choice how technology impacts us.

The Social Age

Aral brings a deep background in social media to this book. He is the David Austin Professor of Management, Marketing, IT and Data Science at MIT, the director of the MIT Initiative on Digital Economy and a self-professed "data nerd." Throughout his career, he has worked with many of the social media platforms he now scrutinizes.

He quickly sets the tone in the introduction of the book: "over the last decade, we’ve doused our kindling fire of human interaction with high-octane gasoline. We’ve constructed an expansive, multifaceted machine that spans the globe and conducts the flow of information, opinions, and behaviors through society." The machine's objective? To change how we think.

Aral describes this system as a self-feeding organism, one that generates data by tracing "each of our preferences, desires, interests, and time-stamped, geolocated activities around the world" and then using that same data to improve its analysis and persuasiveness. As a marketer, I value this component for creating awareness, but the risks of it being applied for no good are also always present.

The interesting thing is the “more precise it (the hype machine) gets, the more engaging and persuasive it becomes.” With its growing powers of persuasion, comes growth in revenue and reach. Social signals now bombard us, "with streams of status updates, news stories, tweets, pokes, posts, referrals, advertisements, notifications, shares, check-ins, and ratings from peers in our social networks, news media, advertisers, and the crowd." Together these signals are remaking the human social network and the flow of information through it. What this also unleashes are “nation-states, businesses, and individuals eager to steer the global conversation toward their ends, to mold public opinion, and ultimately to change what we do.” The urgency of Aral's message has only increased during the COVID-19 pandemic, when stay at home orders pushed even more people onto social media to keep in touch.

Related Article: Tim Berners-Lee Has a Plan to Reinvent the Web

Why Does It Matter?

You might be asking now why should I care? There is no privacy and after all, I'm a free thinker.

The problem Aral uncovers is we're not as impervious to the sways of social media as we'd like to believe. Aral has spent several years raising the issue of "fake news" and been recognized for his efforts in doing so. With "The Hype Machine" he shares exactly how insidious it can be: "false news spreads significantly farther, faster, deeper, and more broadly than the truth — sometimes by an order of magnitude. While the truth rarely diffused to more than 1,000 people, the top 1 percent of false news cascades routinely diffused to as many as 100,000 people." We need look no further than the impact social media has had on elections in recent years for corroboration. As Aral notes, "false political news travels deeper and more broadly, reached more people, and was more viral than any other category of false news. It reached more than 20,000 people nearly three times faster than all other types of false news reached just 10,000 people."

The power of false news transcends follower size, reach, authentication and more. In his research he found, "People who spread false news have significantly fewer followers, followed significantly fewer people, were significantly less active on Twitter, were verified significantly less often, and had been on Twitter for significantly less time, on average." The reach is a result of what Aral calls a "complex interaction of coordinated bots and unwitting humans working together in an unexpected symbiosis."

Related Article: The Next Frontier for IT: AI Ethics

How Does False News Spread so Easily?

The whole approach works when influential people are fooled into sharing false news and content. It starts when fake news is published online.  According to Aral, "social bots [software-controlled social media profiles] pounce on fake news in the first few seconds after it’s published, and they retweet it broadly .... The early tweeting activity by bots triggers a disproportionate amount of human engagement, creating cascades of fake news triggered by bots but propagated by humans through the Hype Machine’s network."

While bots may initiate the sharing of fake news, the spread can't happen without human intervention. One explanation for why humans so willingly play their part here comes from what Soroush Vosoughi, Deb Roy, and Aral call the novelty hypothesis. Humans are attracted to novelty because it is surprising and arouses an emotional reaction. Combine this proclivity towards novelty with the segmentation by bias and prejudices the social platform algorithms break us into and you've got a situation primed for the spread of false news.

As Aral writes, "Attention is valuable because it’s a precursor to persuasion. Platforms like Facebook, Twitter, and YouTube provide connections, communication, and content to get consumer’s attention. They then sell that attention to brands, governments, and politicians who want to change people’s perceptions, opinions, and behavior with ads."

While this ability to micro-target people is a boon to marketers, in the end, what makes social media effective at marketing makes it effective at manipulating and attack from outside.

Related Article: A New Social Contract

The Hype Machine and How it Works

According to Aral, three technologies make the hype machine possible. They are: 

  • Digital social networks + Machine Intelligence
  • The Hype Loop
  • Smartphones 

Together, these create the architecture of the Hype Machine that structures our world today.

Medium: Smart Phones

The medium is largely the smartphone. Smartphones create an always-on environment from which the Hype Machine operates. The smartphone represents "the primary input/output device through which we provide information to and receive information from the Hype Machine."

Substrate: Digital Networks

Digital social networks structure the flow of information and influence society in two ways: They provide access to digital connections and influencing the structure of those connections through friend suggestion algorithms and they use data about our connections to build more accurate models of our preferences. The substrate at the core of the Hype Machine is the network itself: the constantly evolving population-scale collection of links that connects us on Facebook, Twitter, LinkedIn and the rest.

Substrate: Machine Intelligence

Machine intelligence ingests our thoughts, behaviors and opinions and, in turn, curates the stories we see in our newsfeeds, the pictures we see on Instagram, the colleagues and dates suggested to us on LinkedIn and Tinder, and the ads we are shown alongside this content.

To be honest, I had no idea how artificial intelligence and machine learning are used to predict my personal social graph. This book is worth reading to understand the sophistication of social media analytics alone.

The Hype Loop

Aral defines the Hype Loop as "the cyclical interplay of machine and human intelligence that determines what we focus on and how information and knowledge are distributed." Who we are connected to directly affects everything we see online, from the news we read to the products we are pitched. This network has a specific, complex structure that determines who we know and interact with and the pathways through which information, resources and ideas flow to us.

Two regularities of the social graph directly influence what we are experiencing on the Hype Machine today.

First, its meaning emanates from dense clusters of highly connected people. Our connections within these clusters are much stronger than those across clusters. Second, social graphs are homophilous, meaning similar people connect (the so-called birds of a feather flock together). These two properties explain why the Hype Machine helps foster political polarization and echo chambers, spreads fake news, and generates outsize returns on marketing investments.

hype loop

The Hype Loop works through the interplay of machine intelligence and human behavior. The right side of the loop depicts the sense and suggest loop, the process of machine intelligence structuring human choices. The left side depicts the consume and act loop, the process of human agency consuming machine recommendations and acting on them.

We tend to choose the suggested friends, the suggested news stories and the suggested products machine intelligence serves up because we don’t have the time or attention to search more broadly. In some instances, we don’t even see choices because the Hype Machine eliminates them from view. Our behavior — what we post, what we read, how we make friends, and how we communicate and interact with one another — shapes how the Hype Machine interprets what we want from technology and how we want to live. 

The Hype Machine encompasses the technology trifecta above of digital social networks (the substrate), smartphones (the medium), and machine intelligence (the Hype Loop process); the four levers of money, code, norms and laws; and the three trends, including personalized mass persuasion, hyper socialization and the attention economy (the tyranny of trends).

Related Article: Social Media's Dirty Secret

The Message for Marketers 

Aral includes a specific message for marketers: Digital ads don’t work nearly as well as is promised. 

His data did find search ads were the most effective with new and infrequent customers. This outreach to new customers, known as prospecting, spreads brand awareness to people who are unfamiliar with the brand. While Aral has seen brand lift as a result of digital and social media messaging, he warns most companies aren't measuring or managing their efforts properly. So while marketers can benefit from the same effects described above — where attention is the precursor to persuasion — without the proper maintenance, their impact will lessen. 

And again, the ability to direct ads at narrower and narrower categories of consumers comes with a flip side: these same tools can be used to manipulate and attack by others.  

Fixing the Hype Machine

The main motivation behind the book according to Aral "is that we can achieve the promise of social media while avoiding the peril. By looking under the hood at how the Hype Machine operates and employing science to decipher its impact, we can collectively steer this ship away from the impending rocks and into calmer waters."

Unsurprisingly, there's no quick fix here. Aral argues we must collectively use the four levers noted above — money, code, norms and law — to counteract the power of the Hype Machine. He acknowledges this will be particularly at a time when society is at a "crossroads of privacy and security, free speech and hate speech, truth and falsity, democracy and authoritarianism, inclusion and polarization." Fixing this has multiple dimensions. We need consumers to have data portability between social networks as well as enforceable privacy legislation that protects our rights and minimizes harm from data breaches.

Beyond this, Aral believes we need to take concrete actions against fake news and misinformation. These include the following:

  1. Labeling for incorrect information.
  2. Addressing the economic incentives behind the creation and spread of false information.
  3. Media literacy.
  4. Technology solutions to slow the spread of misinformation.
  5. More open collaboration between social media platforms and independent researchers to understand the threats.
  6. Platform policies, such as limiting the number of times a person can forward a message.
  7. Incentivize organizations to make tough content moderation decisions without the fear of civil prosecution.
  8. Eliminate immunity for community-based communication platforms from prosecution when organizations fail to moderate content.

Media literacy is one of the key tactics here, giving people the critical thinking skills to understand the information they read and share. People must be capable of spotting fake news, distinguish fact from opinion and understand the use of media for persuasion. 

A Clear Choice

As I wrote at the beginning of this post, we have a choice with all technology. Weeks after the last election, I met with a leader in data privacy. They told me they had just come from meetings with the various US intelligence agencies. It was clear via the signatures that Russians had actively worked to impact our election. Now that this widely accepted fact, the question is what will do to respond to the hype machine.

As Aral so clearly puts it, the fine art of marketing segmentation and algorithms has broken us into groups that are primed to accept marketing or an enemy state’s message. Our problem, just as with COVID-19, is we can act as super spreaders or we can wear a mask. As such, we all need, regardless of political beliefs, to ensure we do not push information that hasn’t been verified. We must get beyond personal biases and personal prejudices.

The question now is what actions do we want take to protect our citizenship?