The Hype Machine

Copyright

HarperCollinsPublishers

1 London Bridge Street

London SE1 9GF

www.harpercollins.co.uk

First published in the United States by Currency, an imprint of Random House, a division of Penguin Random House LLC, New York.

This edition published by HarpercollinsPublishers 2020

FIRST EDITION

Text © 2020 by HyperAnalytic, Inc.

Cover layout design ©HarperCollinsPublishers Ltd 2020

A catalogue record of this book is available from the British Library

Sinan Aral asserts the moral right to be identified as the author of this work

All rights reserved under International and Pan-American Copyright Conventions. By payment of the required fees, you have been granted the nonexclusive, non-transferable right to access and read the text of this e-book on screen. No part of this text may be reproduced, transmitted, downloaded, decompiled, reverse engineered, or stored in or introduced into any information storage retrieval system, in any form or by any means, whether electronic or mechanical, now known or hereinafter invented, without the express written permission of HarperCollins e-books.

Find out about HarperCollins and the environment at www.harpercollins.co.uk/green

Source ISBN 9780008277116

Ebook Edition © September 2020 ISBN: 9780008277123

Version 2020-09-04

Note to Reader

This ebook contains the following accessibility features which, if supported by your device, can be accessed via your ereader/accessibility settings:

 Change of font size and line height

 Change of background and font colours

 Change of font

 Change justification

 Text to speech

 Page numbers taken from the following print edition: ISBN 9780008277116

Dedication

To my parents—Thank you for every opportunity,

every conversation, and every hug.

To Kaya—May you always stay curious, creative,

and hungry for knowledge.

Contents

COVER

TITLE PAGE

COPYRIGHT

NOTE TO READER

DEDICATION

PREFACE | Pandemics, Promise, and Peril

CHAPTER 1 | The New Social Age

CHAPTER 2 | The End of Reality

CHAPTER 3 | The Hype Machine

CHAPTER 4 | Your Brain on Social Media

CHAPTER 5 | A Network’s Gravity Is Proportional to Its Mass

CHAPTER 6 | Personalized Mass Persuasion

CHAPTER 7 | Hypersocialization

CHAPTER 8 | Strategies for a Hypersocialized World

CHAPTER 9 | The Attention Economy and the Tyranny of Trends

CHAPTER 10 | The Wisdom and Madness of Crowds

CHAPTER 11 | Social Media’s Promise Is Also Its Peril

CHAPTER 12 | Building a Better Hype Machine

NOTES

LIST OF SEARCHABLE TERMS

ACKNOWLEDGMENTS

ILLUSTRATION SOURCES

ABOUT THE AUTHOR

PRAISE FOR SINAN ARAL AND THE HYPE MACHINE

ABOUT THE PUBLISHER

PREFACE

Pandemics, Promise, and Peril

The coronavirus pandemic of 2020 was a “black swan” event, the repercussions of which were felt throughout the world’s health systems, economy, and the very fabric of everyday life. Everyone on earth remembers where they quarantined, who they longed to see, and how they coped with the tremendous mental and physical strain the virus exacted. But another dramatic, albeit subtler, consequence of COVID-19 was the rather abrupt shock it delivered to the world’s global communication system—the central nervous system of digital connections that links our planet. Times Square, Trafalgar Square, and Tahrir Square became ghost towns. And as the virus sent humanity scurrying off the streets and into their homes, it pushed billions of people onto their laptops and smartphones, scrambling to get online. The world logged on to Facebook, Twitter, WhatsApp, Instagram, YouTube, and LinkedIn in record numbers, desperate for news, medical information, social support, human connection, and jobs. The day the offline world stood still, the online world ignited like a digital forest fire.

Demand for social media skyrocketed. Facebook Messenger, WhatsApp, and Facebook Live saw 50 percent increases in usage overnight. Voice calling over Facebook’s apps doubled, while group calls in Italy grew by over 1,000 percent. As movie theaters closed, new downloads of Netflix jumped 66 percent in Italy and 35 percent in Spain. Netflix crashed under the weight of the surge. YouTube was forced to throttle its video quality to handle the deluge. And the Internet stuttered repeatedly throughout the pandemic.

Social collaboration tools exploded with activity. Slack CEO Stewart Butterfield tweeted, “1,597 days after hitting 1M *simultaneously connected* users in Oct ’15 we pass ten million. 6 days later: 10.5M, then 11.0M. Next day, 11.5M. This Monday, 12M. Today 12.5M.” He attached a graph showing the number of “newly created [Slack] work teams” to his tweet thread, which, after March 12, looked like a hockey stick with an extra-long handle pointing straight up. Digital natives were already on social media, but the coronavirus forced many digital Luddites to use social technologies for the first time. New users flocked to social platforms in droves, building armies of fresh profiles, wiring into what I call the “Hype Machine”—the real-time communications ecosystem created by social media. Describing their attempts to cope with the demand, Alex Schultz and Jay Parikh, Facebook’s heads of analytics and engineering, respectively, wrote, “Usage growth from COVID-19 is unprecedented across the industry, and we are experiencing new records in usage every day.” Mark Zuckerberg was blunter: “We’re just trying to keep the lights on over here,” he said.

As the entire planet was denied physical contact for months on end, the coronavirus shocked our use and perception of social technologies in dramatic ways. Facebook, Twitter, WhatsApp, and Instagram became indispensable sources of human connection, timely medical information, social support, outreach, pandemic fundraising, free impromptu concerts, collaborative art projects, and real-time updates about the spread of the virus. Families played Monopoly over Facebook, friends attended cocktail mixers in live group chats, neighbors maintained vibrant WhatsApp groups, and many followed the news on Twitter. Social technologies connected humanity when the planet was forced to disconnect. Group video kept families together. Daily messages kept tabs on our parents and our kids. Hangouts kept work teams collaborating while the world stood still. Even my six-year-old, who gets almost no screen time, connected with his friends and his first-grade class every day over the Hype Machine.

Social platforms provided access to critical medical information about how to socially distance, whether to wear masks, where the hot spots were, and how to be safe at home. The platform companies got right to work, providing new services and data to model and mitigate the spread of the pandemic. Facebook used aggregate anonymous mobility data to create “colocation” maps of where people from different geographies were most likely to cross paths, which aided epidemiologists modeling where the pandemic was likely to spread next, from one geography to another. They spun up “disease prevention maps” that leveraged Facebook’s mobile app data to help understand how people were moving and therefore how they might be spreading the virus.

As the director of the MIT Initiative on the Digital Economy, one of the largest centers researching the influence of digital technologies on our world, and head of MIT’s Social Analytics Lab, a team of thirty brilliant MIT faculty and students studying the impact of social media on society, I felt we needed “all hands on deck” to address the crisis. We organized Zoom meetings to brainstorm how we could contribute to efforts to address the pandemic. While stressing the need to first do no harm, I reached out to my contacts at social platforms around the world to see how we could help.

In a week, we created three projects supporting national and international health organizations, measuring the effect of social distancing on COVID-19 spread and fighting pandemic misinformation online. I reached out to Facebook and suggested a collaboration. They responded quickly and said that, since we already had a data licensing agreement, they could share data right away. We focused on modeling the effect of social distancing on the pandemic’s spread. Facebook’s disease prevention maps tracked aggregated, anonymous Facebook mobile app data on user location density, movement, and network connectivity, as well as the average number of half-kilometer-squared spaces in which Facebook users were present each day, compared to precrisis levels. We worked to combine this data with detailed records of the social distancing orders imposed on different regions, states, and cities around the world. The idea was to estimate the effect of social distancing orders on the number of Facebook users present in public places within the region or city in which the order was imposed and how those effects spilled over into other regions. We wanted to know if distancing orders worked and, more important, when they didn’t and why. We also advised Facebook’s COVID Symptom Survey, which supplemented traditional disease surveillance with a million surveys per week in the United States, asking Facebook users if they had common COVID symptoms and whether they were staying home. As I write this, Facebook is preparing to launch a global survey to users worldwide.

I then called John Kelly and Camille François of Graphika, one of two companies commissioned by the U.S. Senate Intelligence Committee to investigate Russian interference in the 2016 and 2020 U.S. presidential elections, to brainstorm how we could track and fight the spread of COVID misinformation and the growing threat of election misinformation ahead of the 2020 election. We decided to monitor, track, and publicly report on the automated software “bot,” cyborg, and troll networks spreading misinformation about coronavirus and the election around the world, and to measure how misinformation was affecting disease prevention and voting.

I also connected with Gustav Praekelt of the Praekelt Foundation in South Africa. Gustav and I already had an active, countrywide project using social media to fight the spread of HIV/AIDS in South Africa. When I asked him how we could collaborate on COVID-19, he told me he had repurposed the WhatsApp and Messenger tools we were using in our HIV project to spread official COVID medical information around the world. They had created something called COVIDConnect, an automated software robot that would field questions from the public and respond to them with the correct information from official sources over WhatsApp, Facebook Messenger, and text messaging. COVIDConnect had become the official engine of the global World Health Organization (WHO) WhatsApp channel and powered the automated national COVID response hotlines in South Africa, New Zealand, Australia, and ten other countries in Africa and Southeast Asia, amassing 15 million users in just two weeks. It was fast becoming the go-to source for official information about COVID. But Praekelt and Facebook had a problem. They were worried about the spread of COVID misinformation on WhatsApp. It’s hard to root out misinformation on an encrypted platform like WhatsApp, because messages can’t be publicly tracked. So we got to work building a system to debunk COVID misinformation over the official WHO and national WhatsApp accounts.

The social platforms addressed the economic fallout from the pandemic as well. Small businesses used their Facebook pages to sell online. Live videocasts were hosted across social media to replace in-store events that usually generate foot traffic and boost sales. Stage shows were produced and aired over Instagram Stories and TikTok. Yoga classes, guitar lessons, and hairstylist sessions all transitioned to the Hype Machine. Facebook even set up a $100 million small-business relief fund to dole out no-strings-attached cash grants to keep small businesses afloat.

These projects were just beginning as I finished this book. The people who make the social media industry tick are dedicated technologists. They care about the future of our planet. They’re wicked smart, and they’re committed to making the world a better place. But social media’s impact on the world is not determined by intention alone. As we all know, there have been many missteps in building the Hype Machine.

Following the death of George Floyd, Mark Zuckerberg defended his decision to allow unaltered and unlabeled, divisive and inflammatory Facebook messages by President Trump that seemed to threaten violence in response to the protests and including the words “when the looting starts the shooting starts,” a phrase used by police chiefs and segregationist politicians during crackdowns against the civil rights movement. Twitter limited the public’s exposure to the president’s messages, saying it “violated the Twitter Rules about glorifying violence.” Zuckerberg’s inaction led to virtual walkouts by Facebook employees unable to align themselves with a company policy that seemed to appease racist rhetoric.

The backlash against Zuckerberg was a return to an old refrain. Our enthusiastic embrace of social media during the pandemic lockdown of 2020 was a 180-degree reversal from 2019, in the weeks and months before COVID hit. Before the pandemic, social media was a pariah. The #deletefacebook movement was gaining steam. The Cambridge Analytica scandal had forced Mark Zuckerberg to testify on Capitol Hill and in front of the European Parliament. Lawmakers were angling to break up the social media giants on antitrust grounds. Just weeks before the virus hit China, Sacha Baron Cohen called social media “the greatest propaganda machine in history.” In his remarks before the Anti-Defamation League, he said, “Facebook, Google, YouTube, Twitter, and others … reach billions of people. The algorithms these platforms depend on deliberately amplify … stories that trigger outrage and fear. … It’s why fake news outperforms real news, because studies show lies spread faster than the truth.” The study he was referring to was one I published with my colleagues Deb Roy and Soroush Vosoughi in a cover story for Science, called “The Spread of True and False News Online.” I’ll return to it in the first two chapters of The Hype Machine.

But misinformation wasn’t the only concern of social media critics before the pandemic. Facebook, Twitter, WhatsApp, Instagram, and YouTube were undermining our privacy, facilitating foreign interference in our democracy, threatening the integrity of our elections, promoting political lies for money, radicalizing terrorists, chilling free speech, and promoting hate speech. They spread genocidal propaganda, livestreamed mass murders like the one in Christchurch, New Zealand; promoted predatory lending to minorities; discriminated against women in employment advertising; tracked our every move; manipulated our emotions for revenue; and promoted political polarization.

Social media’s promise and peril even played out in debates about the pandemic itself. Yes, we relied on social media for human connection, social support, and lifesaving information. But it was also a cauldron of misinformation about impending national lockdowns and false cures, nationalistic finger-pointing between the United States and China, and foreign interference designed to fan the flames of our fears. Privacy debates took on new meaning during the COVID crisis as the threat of “surveillance capitalism” morphed into lifesaving “disease surveillance.” Facebook wasn’t surveilling for profit during COVID; it was filling gaps in inadequate national disease surveillance programs with scalable symptom surveys that identified the pandemic’s spread. At the same time, Google, Apple, and MIT developed Bluetooth-based contact tracing systems that would alert users who opted in if they had come in close physical proximity to the Bluetooth-enabled device of a COVID carrier. Privacy advocates listened in horror as the tech giants swore the system would remain anonymous. The perils of social media surveillance for privacy were sharply contrasted with the promise of social media surveillance for health. Some thought the surveillance was worth the privacy risk. Others thought the risks outweighed the benefits.

So which of these visions of social media is correct—the promise or the peril? Is the Hype Machine a force for good, for collective intelligence and solidarity? Or is it a scourge and a pariah? The global reach of Facebook, WhatsApp, Twitter, Instagram, and YouTube affects our communications, data, and privacy, as well as the flow of information around the world. It has the potential to be uniquely dangerous. But the pandemic reminded us how invaluable this far-reaching global communication network is to us, especially in times of need, and just how much we’ve come to rely on it for everything from human connection to news to job opportunities to entertainment to dating and relationships in our everyday lives.

During the pandemic, the source of the Hype Machine’s promise was also the source of its peril. As I will describe, this is true of social media in general, making it difficult to regulate. The solutions that social media enable compromise our privacy beyond what some think is necessary. But, in Europe for example, stringent privacy regulations prevented the use of social media to track and tackle the pandemic. These examples highlight many important unanswered questions.

Is social media a force for meaningful connection, collaboration, social support, and access to lifesaving information? Or is it a propaganda machine that, left unchecked, will destroy democracy, civil society, and our health? Can the promise of social media be realized without the peril? Or are they inexorably linked?

As I explore in this book, the Hype Machine has the potential for both promise and peril. And the decisions we make in the next eighteen to twenty-four months in how we design, regulate, monetize, and use social media will determine which path we realize. We’re at a crossroads. To act responsibly, we have to educate ourselves about how social media works.

My goal is to take you on a roller-coaster journey through what I’ve learned studying, building, investing in, and working with social media over the last twenty years. It’s a harrowing journey with unbelievable discoveries and sordid scandals about how social media impacts our democracy; how it can disseminate lies while connecting us to valuable truths; how it fights repression at times while promoting it at others; how it propagates hate speech while defending free speech; and, most of all, how all this works, under the hood, to hook us neurologically, emotionally, socially, and economically. The story not only reveals the business strategies behind social media, but also the relationship between social media’s design and how it affects us.

The story of the Hype Machine as either the promise or the peril alone is one-sided. The truth is more complicated—it’s sometimes uplifting and at other times depressing, sometimes shocking but always illuminating. Most of all, it’s a story that involves every one of us, now and for generations to come.

CHAPTER 1

The New Social Age

This is the whole point of technology. It creates an appetite for immortality on the one hand. It threatens universal extinction on the other. Technology is lust removed from nature.

—DON DELILLO

Human beings have always been a social species. We’ve been communicating, cooperating, and coordinating with one another since we were hunting and gathering. But today something is different. 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. This Hype Machine connects us in a worldwide communication network, exchanging trillions of messages a day, guided by algorithms, designed to inform, persuade, entertain, and manipulate us.

The object of this machine is the human psyche. It was designed to stimulate our neurological impulses, to draw us in and persuade us to change how we shop, vote, and exercise, and even who we love. It analyzes us to give us options for what to read, buy, and believe. It then learns from our choices and iteratively optimizes its offerings. As it operates, it generates a data exhaust that traces each of our preferences, desires, interests, and time-stamped, geolocated activities around the world. It then feeds on its own data exhaust, refining its process, perfecting its analysis, and improving its persuasive leverage. Its motivation is money, which it maximizes by engaging us. The more precise it gets, the more engaging and persuasive it becomes. The more persuasive it becomes, the more revenue it generates and the bigger it grows. This is the story of the Hype Machine—the social media industrial complex: how it was designed, how it works, how it affects us, and how we can adapt to it. And the story opens in Crimea.

Ten Days

On a cold day in February 2014, heavily armed gunmen surrounded the Crimean parliament building in Simferopol, Ukraine. They wore no sovereign markings but were later confirmed to be Russian special forces reacting to the deposition of Ukrainian president Viktor Yanukovych just days before. By all accounts, the gunmen were organized and professional. After breaking through the front door, they cut the building’s communications, confiscated all mobile electronic devices, and systematically controlled who entered and exited the building, maintaining a tight perimeter and allowing no foreign journalists inside.