Most social-media platforms rely on advertising. Advertisers pay for engagement, such as clicks, likes and shares. Therefore, sensational and attention-grabbing content is more valuable. This motivates platforms to use automated-recommendation technology that relies on algorithmic decision-making to prioritize content likely to grab attention. The main strategy of “user-targeted amplification” shows users content that is most likely to interest them based on detailed data that are collected about them. See more in the Risk section under Data Monetization by social media companies and tailored information streams.

The Emergence of Programmatic Advertising

The transition of advertising to digital systems has dramatically altered the advertising business. In an analog world, advertising placements were predicated on aggregate demographics, collected by publishers and measurement firms. These measurements were rough, capable at best of tracking subscribers and household-level engagement. Advertisers hoped their ads would be seen by enough of their target demographic (for example, men between 18 and 35 with income at a certain level) to be worth their while. Even more challenging was tracking the efficacy of the ads. Systems for measuring if an ad resulted in a sale were limited largely to mail-in cards and special discount codes.

The emergence of digital systems changed all of that. Pioneered for the most part by Google and then supercharged by Facebook in the early years of the 21st century, a new promise emerged: “Place ads through our platform, and we can put the right ad in front of the right person at the right time. Not only that, but we can report back to you (advertiser) who saw the ad, if they clicked on it, and if that click led to a ‘conversion’ or a sale.”

But this promise has come with significant unintended consequences. The way that the platforms—and the massive ad tech industry that has rapidly emerged alongside them—deliver on this promise requires a level of data gathering, tracking and individual surveillance unprecedented in human history. The tracking of individual behaviors, preferences and habits powers the wildly profitable digital advertising industry, dominated by platforms that can control these data at scale.

Managing huge consumer data sets at the scale and speed required to deliver value to advertisers has come to mean a heavy dependence on algorithms to do the searching, sorting, tracking, placement and delivery of ads. This development of sophisticated algorithms led to the emergence of programmatic advertising, which is the placement of ads in real time on websites with no human intervention. Programmatic advertising made up roughly two thirds of the $237 billion global ad market in 2019.

The digitization of the advertising market, particularly the dominance of programmatic advertising, has resulted in a highly uneven playing field. The technology companies enter with a significant advantage: they built the new structures and set the terms of engagement. What began as a value add in the new digital space— “We will give advertisers efficiency and publishers new audience and revenue streams”—has evolved to disadvantage both.

One of the primary challenges is in how audience engagement is measured and tracked. The primary performance indicators in the digital world are views and clicks. As mentioned above (and well documented in the literature), an incentive structure based on views and clicks (engagement) tends to favor sensational and eye-catching content. In the race for engagement, misleading or false content, with dramatic headlines and incendiary claims, consistently wins out over more balanced news and information. See also the section on digital advertising in the disinformation resource.

Advertising motivated content

Platforms leverage tools like hashtags and search engine optimization (SEO) to rank and cluster content around certain topics. Unfortunately, automated content curation motivated by advertising does not tend to prioritize healthful, educational, or rigorous content. Instead, frivolous, distracting, potentially untrue or even harmful content tends to spread more widely: conspiracy theories, shocking or violent content and “click-bait” (misleading phrases designed to entice viewing). Many platforms have features of upvoting (“like” buttons) which, similar to hashtags and SEO, influence the algorithmic moderation and promote certain content to circulate more widely. These features together cause “virality,” one of the defining features of the social-media ecosystem: the tendency of an image, video, or piece of information to be circulated rapidly and widely.

In some cases, virality can spark political activism and raise awareness (like the #MeToo hashtag), but it can also amplify tragedies and spread inaccurate information (anti-vaccine information and other health rumors, etc.). Additionally, the business models of the platforms reward quantity over quality (number of “likes”, “followers”, and views), encouraging a growth logic that has led to the problem of information saturation or information overload, overwhelming users with seemingly infinite content. Indeed, design decisions like the “infinite scroll” intended to make our social media spaces ever larger and more entertaining have been associated with impulsive behaviors, increased distraction, attention-seeking behavior, lower self-esteem, etc.

Many digital advertising strategies raise risks regarding access to information, privacy, and discrimination, in part because of their pervasiveness and subtlety. Influencer marketing, for example, is the practice of sponsoring a social media influencer to promote or use a certain product by working it into their social-media content, while native advertising is the practice of embedding ads in or beside other non-paid content. Most consumers do not know what native advertising is and may not even know when they are being delivered ads.

It is not new for brands to strategically place their content. However, today there is much more advertising, and it is seamlessly integrated with other content. In addition, the design of platforms makes content from diverse sources—advertisers and news agencies, experts and amateurs—indistinguishable. Individuals’ right to information and basic guarantees of transparency are at stake if advertisements are placed on equal footing with desired content.

Content moderation is at the heart of the service that social-media platforms provide: the hosting and curation of the content uploaded by their users. Content moderation is not just the review of content, but every design decision made by the platforms, from the Terms of Service and their Community Guidelines, to the algorithms they use to rank and order content, to the types of content they allow and encourage through design features (“like”, “follow”, “block”, “restrict”, etc.).

Content moderation is particularly challenging because of the issues it raises around freedom of expression. While it is necessary to address massive quantities of harmful content that circulate widely, educational, historic, or journalistic content is often censored by algorithmic moderation systems. In 2016, Facebook took down a post with a Pulitzer Prize-winning image of a naked 9-year-old girl fleeing a napalm bombing and suspended the account of the journalist who had posted it.

Though nations differ in their stances on freedom of speech, international human rights provide a framework for how to balance freedom of expression against other rights, and against protections for vulnerable groups. Still, content-moderation challenges increase as content itself evolves, for instance through increase of live streaming, ephemeral content, voice assistants, etc. Moderating internet memes is particularly challenging, for instance, because of their ambiguity and ever-changing nature; and yet meme culture is a central tool used by the far right to share ideology and glorify violence. Some communications manipulations are also intentionally difficult to detect, for example, “dog whistling” (sending coded messages to subgroups of the population) and “gaslighting” (psychological manipulation to make people doubt their own knowledge or judgement).

Automated moderation

Content moderation is usually performed by a mix of humans and artificial intelligence , with the precise mix dependent on the platform and the category of content. The largest platforms like Facebook and YouTube use automated tools to filter content as it is uploaded. Facebook, for example, claims it is able to detect up to 80% of hate speech content in some languages as it is posted, then submitting it for human review. Though the working conditions for the human moderators have been heavily critiqued, the accuracy and transparency of the algorithms are also disputed and, expectedly, have some concerning biases. Humans are of course subject to biases as well, but algorithmic bias in content moderation poses more serious threats around equity and freedom of expression.

The complexity of content-moderation decisions does not lend itself easily to automation, and the porosity between legal and illegal/permissible and impermissible content leads to both cases of legitimate content being censored, and cases of harmful and illegal content passing through the filter (cyberbullying, defamation, etc.).

The moderation of content posted to social media has been increasingly important during the COVID-19 pandemic, when access to misleading and inaccurate information about the virus can result in death. The current moderation strategy of Facebook, for example, has been described as creating “a platform that is effectively at war with itself: the News Feed algorithm relentlessly promotes irresistible click-bait about Bill Gates, vaccines, and hydroxychloroquine; the trust and safety team then dutifully counters it with bolded, underlined doses of reality.”

Addressing harmful content

In some countries, local laws may address content moderation, but they relate mainly to child abuse images or illegal content that incites violence. Most platforms also have community standards or safety and security policies that state the kind of content allowed, and that sets the rules for harmful content. Enforcement of legal requirements and the platforms’ own standards relies primarily on content being flagged by social media users. The social-media platforms are only responsible for harmful content shared on their platforms once it has been reported to them.

Some platforms have established mechanisms that allow civil society organizations (CSOs) to contribute to the flagging process by becoming a so-called trusted flagger. With Facebook, this allows verification of accounts for civic organizations, provides higher levels of protection, faster response for incident reports, and accounts are less easily automatically disabled. For example, Access Now’s Digital Security Helpline is a trusted partner, and Facebook also offers access to its Trusted Partners Program to partners of members of the Design 4 Democracy Coalition. However, this program does not compensate for the limited accessibility to the platform for CSOs that encounter problems.

Social media has been credited with providing channels for citizens, activists, and experts to report instantly and directly—from disaster settings, during protests, from within local communities, etc. Citizen journalism, also referred to as participatory journalism or guerrilla journalism, does not have a definite set of principles and should not be considered as a replacement for professional journalism, but it is an important supplement to mainstream journalism. Collaborative journalism, the partnership between citizen and professional journalists, as well as crowdsourcing strategies, are further techniques permitted by social media that have enhanced journalism, helping to promote voices from the ground and to magnify diverse voices and viewpoints. The outlet France 24 has developed a network of 5,000 contributors, the “observateurs,” who are able to cover important events directly by virtue of being on scene at the time, as well as to confirm the accuracy of information.

Social-media platforms as well as blogging tools have allowed for the decentralization of expertise, bridging elite and non-elite forms of knowledge. Without proper fact-checking or supplementary sources and proper context, citizen reporting carries risks— including security risks to the authors themselves—but it is an important democratizing force and source of information.

In crowdsourcing, the public is mobilized to share data together to tell a larger story or accomplish a greater goal. Crowdsourcing can be a method for financing, for journalism/reporting, or simply for gathering ideas. Usually some kind of software tool or platform is put in place that the public can easily access and contribute to. Crisis mapping, for example, is a type of crowdsourcing through which the public shares data in real time during a crisis (a natural disaster, an election, a protest, etc.). These data are then ordered and displayed in a useful way. For instance, crisis mapping can be used in the wake of an earthquake to show first responders the areas that have been hit and need immediate assistance. Ushahidi is an open-source crisis-mapping software developed in Kenya after the violent outbreak following the election in 2007. The tool was first created to allow Kenyans to flag incidents, to form a complete and accurate picture of the situation on the ground, to share with the media, outside governments, and relevant civil society and relief organizations. In Kenya, the tool gathered texts, tweets, and photos and created crowdsourced maps of incidents of violence, election fraud, and other abuse. Ushahidi now has a global team and works in 30 different languages.

Social media has allowed local and global movements to spring up overnight, inviting broad participation and visibility. Twitter hashtags in particular have been instrumental for coalition building, coordination, and for raising awareness among international audiences, media and government. Researchers began to take note of digital activism around the 2011 “Arab Spring,” when movements in Tunisia, Morocco, Syria, Libya, Egypt and Bahrain, among others countries, leveraged social media and were quickly followed by the Occupy Wallstreet movement in the United States. Ukranian’s Euromaidan movement in late 2013 and the Hong Kong protests in 2019 are also examples of political movements that used social media to galvanize support.

In 2013, the acquittal of George Zimmerman in the death of unarmed 17-year-old Trayvon Martin inspired the creation of the  #BlackLivesMatter hashtag. This movement grew stronger in response to the tragic killing of Michael Brown. The hashtag, at the front of an organized national protest movement, provided an outlet for people to join an online conversation and articulate alternative narratives in real time about subjects that the media and the rest of the United States (and more recently with the killing of George Floyd, the world) had not paid sufficient attention to: police brutality, systemic racism, racial profiling, inequality, etc.

The #MeToo movement against sexual misconduct in the media industry, which also became a global movement, has allowed a multitude of people to participate in activism previously bound to a certain time and place.

Some researchers and activists fear “slacktivism” and the effect of social media giving people an excuse to stay at home rather than make a more dynamic response, and some fear that the tools of social media are ultimately insufficient for enacting meaningful social change, which requires nuanced political arguments. (Interestingly, a 2018 Pew Research survey on attitudes toward digital activism showed that just 39% of white Americans believed social media was an important tool to use to express themselves, while 54% percent of Black people said that it was an important tool for them.)

Social media has enabled new online groups to gather together and to express a common sentiment as a form of solidarity or as a means to protest. Especially since the COVID-19 pandemic broke out, many physical protests have been suspended or cancelled, and virtual protests proceeded.

Social media provides a valuable opportunity for CSOs to reach their goals and engage with existing and new audiences. A good social-media strategy is best underpinned by a permanent staff position to grow a strong and consistent social-media presence based on the organization’s purpose, values, and culture. This person should know how to seek information, be aware of both the risks and benefits of sharing information online and understand the importance of using sound judgment when posting. The USAID ‘Social Networking: A Guide to Strengthening Civil Society through Social Media’ provides a set of questions as guidance to develop a sound social-media policy, asking organizations to think about values, roles, content, tone, controversy and privacy.

Social media can be integrated into programmatic activities to strengthen the reach and impact of the program, for example, by generating awareness of an organization’s services to a new demographic. Organizations can promote their programs and services while responding to questions and fostering open dialogue. Widely used social media platforms can be useful to reach new audiences for training and consulting activities through webinars or individual meetings designed for NGOs.

Social-media fundraising presents an important opportunity for non-profits, but organizations should carefully consider the type of campaign and platforms they choose. TechSoup, a non-profit providing tech support for NGOs, offers advice and an online course on fundraising with social media for non-profits.

After the blast in Beirut’s harbor in the summer of 2020, many Lebanese people started online fundraising pages for their organizations. Social-media platforms were used extensively to share funding suggestions to the global audience watching the disaster unfold, reinforced by traditional media coverage.

In some contexts, civic actors rely on social media platforms to produce and disseminate critical information, for example, during humanitarian crises or emergencies. Even in a widespread disaster, the internet often remains a significant communication channel, which makes social media a useful complementary means for emergency teams and the public. Reliance on the internet, however, increases vulnerability in case of network shutdowns.

The ways in which content flows and is presented in social media due to the platforms’ business models risk limiting our access to information, particularly to information that challenges our preexisting beliefs, by exposing us to content likely to attract our attention and support our views. The concept of the filter bubble refers to the filtering of information by online platforms and through our own intellectual biases that worsen polarization by allowing us to live in echo chambers. This is easily witnessed in a YouTube feed: when you search for a song by an artist, you will likely find more songs by the same artist, or similar ones—the algorithms are designed to prolong your viewing, and assume you want more of something similar. The same has been identified for political content. Social media algorithms encourage confirmation bias, exposing us to content we will agree with and enjoy, often at the expense of the accuracy, rigor, educational or social value of that content.

The massive and precise data amassed by advertisers and social media companies about our preferences and opinions permit the practice of micro-targeting, which involves the display of tailored content based on data about users’ online behaviors, connections, and demographics, among others, as will be further explained below.

The increasingly tailored distribution of news and information is a threat to political discourse, diversity of opinions and democracy. Users can become detached even from factual information that disagrees with their viewpoints and isolated within their own cultural or ideological bubbles.

Because tailoring news and other information on social media is driven largely by nontransparent, opaque algorithms that are owned by private companies, it is hard for users to avoid these bubbles. Access to and intake of the very diverse information available on social media, with its many viewpoints, perspectives, ideas and opinions, requires an explicit effort by the individual user to go beyond passive consumption of the content that is presented to them.

The internet and the dominant online platforms provide new tools that amplify and alter the danger presented by false, inaccurate or taken out of context information. The online space increasingly drives discourse and is where much of today’s disinformation takes root. Refer to the Disinformation resource for a detailed overview of these problems.

Social media facilitates a number of violent behaviors such as defamation, harassment, bullying, stalking, “trolling” and “doxxing.” Cyberbullying among children, like traditional offline bullying, can harm students’ performance in school and causes real psychological damage. Cyberbullying is particularly harmful because victims experience the violence alone, isolated in cyberspace. They often do not seek help from parents and teachers, who they believe are not able to intervene. Cyberbullying is also difficult to address because it can move across social-media platforms, beginning on one and moving to another. Like cyberbullying, cyber harassment and cyberstalking have very tangible offline effects. Women are most often the victims of cyber harassment and cyberviolence, sometimes through the use of stalkerware installed by their partners to track their movements. A frightening cyber harassment trend has accelerated in France during the COVID-19 confinement in the form of “fisha” accounts, where bullies, aggressors or jilted ex-boyfriends will publish and circulate naked photos of teenage girls without their consent.

Journalists, women in particular, are often subject to cyber harassment and threats, particularly those who write about socially sensitive or political topics. Online violence against journalists can lead to journalistic self-censorship, affecting the quality of the information environment and democratic debate. Online tools provide new ways to spread and amplify hate speech and harassment. The use of fake accounts, bots, and even bot-nets (automated networks of accounts) allow perpetrators to attack, overwhelm, and even disable the social media accounts of their victims. Doxxing, by revealing sensitive information about journalists, is another strategy that can be used for censorship.

The 2014 case of Gamergate, when several women video-game developers were attacked by a coordinated harassment campaign that included doxxing and threats of rape and death, illustrates the strength and capacity of loosely-connected hate groups online to rally together, inflict real violence, and even drown out criticism. Many of the actions of the most active Gamergate trolls are illegal, but their identities are unknown. Importantly, it has been suggested by supporters of Gamergate that the most violent trolls were a “smaller, but vocal minority”—evidence of the magnifying power of internet channels and their use for coordinated online harassment.

Online hoaxes, scams, and frauds, like in their traditional offline forms, usually aim to extract money or sensitive information from the target. The practice of phishing is increasingly common on social media: an attacker pretends to be a contact or a reputable source in order to send malware or to extract personal information or account credentials. Spearphishing is a targeted phishing attack that leverages information about the recipient and details related to the surrounding circumstances. Coronavirus-related spearphishing attacks have increased steadily during the pandemic.

Most social-media platforms are free to use. Users simply register and then are able to use the platform. Social-media platforms do not receive revenue directly from users, like in a traditional subscription service; rather they generate profit primarily through digital advertising. Digital advertising is based on the collection of users’ data by social-media companies, which allows advertisers to target their ads to specific users and types of users. Social-media platforms monitor their users and build detailed profiles that they sell to advertisers. The data tracked includes information about the user’s connections and behavior on the platform, such as friends, posts, likes, searches, clicks and mouse movements. Data are also extensively collected outside platforms, including information about users’ location, webpages visited, online shopping and banking behavior. Additionally, many companies regularly access the contact book and photos of their users.

In the case of Facebook, this has led to a long-held and widespread conspiracy theory that the company listens to conversations to serve tailored advertisements. No one has ever been able to find clear evidence that this is actually happening. Research has shown that a company like Facebook does not need to listen in to your conversations, because it has the capacity to track you in so many other ways: “Not only does the system know exactly where you are at every moment, it knows who your friends are, what they are interested in, and who you are spending time with. It can track you across all your devices, log call and text metadata on phones, and even watch you write something that you end up deleting and never actually send.”

The massive and precise data amassed by advertisers and social-media companies about our preferences and opinions permit the practice of micro-targeting, that is, displaying targeted advertisements based on what you have recently purchased, searched for or liked. But just as online advertisers can target us with products, political parties can target us with more relevant or personalized messaging. Studies are currently trying to determine the extent to which political micro-targeting is a serious concern for the functioning of democratic elections. The question has also been raised by researchers and digital rights activists as to how micro-targeting may be interfering with our freedom of thought.

The increasingly tailored distribution of news and information is a threat to political discourse, diversity of opinions and democracy. Users can become detached from information that disagrees with their viewpoints and isolated within their own cultural or ideological bubbles. This is easily witnessed in a YouTube feed: when you search for a song by an artist, you will likely be shown more by the same artist, or similar ones—the algorithms are designed to prolong your viewing, and assume you want more of something similar. The same has been identified for ideological content.

The content shared over social media is monitored by governments, who use social media for censorship, control and information manipulation. Many democratic governments are known to engage in extensive social-media monitoring for law enforcement and intelligence-gathering purposes. These practices should be guided by robust legal frameworks to safeguard individuals’ rights online, such as privacy and data-protection laws , but many countries have not yet enacted these types of laws.

There are also many examples of authoritarian governments using personal and other data harvested through social media to intimidate activists, silence opposition, and bring development projects to a halt. The information shared on social media often allows bad actors to build extensive profiles of individuals that enable targeted online and offline attacks, often through social engineering techniques. For example, a phishing email can be carefully crafted based on social-media data to trick an activist to click on a malicious link that provides access to the device, documents or social-media accounts.

Sometimes, however, a strong, real-time presence on social media can protect a prominent activist against threats by the government. A disappearance or arrest will be immediately noticed by followers or friends of a person who suddenly becomes silent on social media.

We rely on social-media platforms to help fulfill our fundamental rights (freedom of expression, assembly, etc.). However, these platforms are massive global monopolies and have been referred to as “the new governors.” This market concentration is troubling to national and international governance mechanisms. Simply breaking up the biggest platform companies will not fully solve the information disorders and social problems presented by social media. Civil society and governments also need visibility of the design choices made by the platforms to understand how to address their negative aspects.

The growing influence of social-media platforms has given many governments reasons to impose laws on online content. There is a surge in laws across the world regulating illegal and harmful content, such as incitement to terrorism or violence, false information, and hate speech. These laws often criminalize speech and contain punishments of jail terms or high fines for something like a retweet on Twitter. Even in countries where the rule of law is respected, legal approaches to regulating online content may be ineffective due to the many technical challenges of content moderation. There is also a risk of violating internet users’ freedom of expression by reinforcing imperfect and non-transparent moderation practices and over-deletion. Lastly, they constitute a challenge to social media companies to navigate between compliance with local laws and defending international human rights law.

As noted above, virality is one of the defining features of the social-media ecosystem: the tendency of an image, video, or piece of information to circulate rapidly and widely. In some cases, virality can spark political activism and raise awareness (like the #MeToo hashtag), but it can also amplify tragedies (the video of the Christchurch massacre in New Zealand) and spread inaccurate information.

Social media has had a profound impact on the field of journalism. While it has allowed the emergence of the citizen-journalist, locally-reported and crowd-sourced information, social-media companies have displaced the relationship between advertising and the traditional newspaper, and created a rewards system for sensationalist, click-bait content that will attract the widest attention globally, over quality journalism that may be pertinent to local communities. In many places, these monopolies have also been partially responsible for the collapse of local news.

The reason for this impact is that, while advertising has successfully made the transition to digital, with global revenues currently at $247 billion and growing by 4% year over year, very little of that revenue is making its way to publishers. The ad tech supply chain, dominated by Google and Facebook, now consumes 90% of all new growth in the world’s major markets and 61 cents of every dollar spent on digital advertising worldwide.

The disruption of the publishing business model has been a slow-motion disaster for news organizations around the world. Since 2012, digital newspaper ad revenues worldwide have grown from an anemic $7.3 billion to $9.95 billion in 2016, while the Google/Facebook duopoly will earn $174 billion or 61% of the global digital advertising market.

In addition, the way search tools work dramatically affects local publishers, as search is a powerful vector for news and information. Researchers have found that search rankings have a marked impact on our attention. Not only do we tend to think information that is ranked more highly is more trusted and relevant, but we tend to click on top results more often than lower ones. The Google search engine concentrates our attention on a narrow range of news sources, a trend that works against diverse and pluralistic media outlets. It also tends to work against the advertising revenue of smaller and community publishers, which is based on user attention and traffic. It is a downward spiral: search results favor larger outlets, those results drive more user engagement which makes their inventory more valuable in the advertising market, those publishers grow larger driving more favorable search results and onward we go.

Amelia Hunt & Doug Specht, Journal of International Humanitarian Action, 2019

“Crowdsourced mapping has become an integral part of humanitarian response, with high profile deployments of platforms following the Haiti and Nepal earthquakes, and the multiple projects initiated during the Ebola outbreak in North West Africa in 2014, being prominent examples. There have also been hundreds of deployments of crowdsourced mapping projects across the globe that did not have a high profile. This paper, through an analysis of 51 mapping deployments between 2010 and 2016, complimented with expert interviews, seeks to explore the organisational structures that create the conditions for effective mapping actions, and the relationship between the commissioning body, often a non-governmental organisation (NGO) and the volunteers who regularly make up the team charged with producing the map.”

Marcia Mundt, Karen Ross , and Charla M Burnett, Social Media and Society, 2018

“Drawing on a case study of Black Lives Matter (BLM) that includes both analysis of public social media accounts and interviews with BLM groups, [the authors] highlight possibilities created by social media for building connections, mobilizing participants and tangible resources, coalition building, and amplifying alternative narratives. [They] also discuss challenges and risks associated with using social media as a platform for scaling up. Our analysis suggests that while benefits of social media use outweigh its risks, careful management of online media platforms is necessary to mitigate concrete, physical risks that social media can create for activists.”

Evelyn Douek, LawFare, 2018

“Members of the Myanmar military have systematically used Facebook as a tool in the government’s campaign of ethnic cleansing against Myanmar’s Rohingya Muslim minority, according to an incredible piece of reporting by the New York Times on Oct. 15. The Times writes that the military harnessed Facebook over a period of years to disseminate hate propaganda, false news and inflammatory posts. The story adds to the horrors known about the ongoing violence in Myanmar, but it also should complicate the ongoing debate about Facebook’s role and responsibility for spreading hate and exacerbating conflict in Myanmar and other developing countries…”

Gary King, Jennifer Pan & Margaret E. Roberts, American Political Science Review, 2017 (Study)

“The Chinese government has long been suspected of hiring as many as 2,000,000 people to surreptitiously insert huge numbers of pseudonymous and other deceptive writings into the stream of real social media posts, as if they were the genuine opinions of ordinary people. Many academics, and most journalists and activists, claim that these so-called “50c party” posts vociferously argue for the government’s side in political and policy debates. As we show, this is also true of the vast majority of posts openly accused on social media of being 50c. Yet, almost no systematic empirical evidence exists for this claim, or, more importantly, for the Chinese regime’s strategic objective in pursuing this activity. In the first large scale empirical analysis of this operation, we show how to identify the secretive authors of these posts, the posts written by them, and their content. We estimate that the government fabricates and posts about 448 million social media comments a year. In contrast to prior claims, we show that the Chinese regime’s strategy is to avoid arguing with skeptics of the party and the government, and to not even discuss controversial issues. We show that the goal of this massive secretive operation is instead to distract the public and change the subject, as most of these posts involve cheerleading for China, the revolutionary history of the Communist Party, or other symbols of the regime. We discuss how these results fit with what is known about the Chinese censorship program and suggest how they may change our broader theoretical understanding of “common knowledge” and information control in authoritarian regimes.”

The World Wide Fund For Nature (WWF) launched Earth Hour in 2007. This social media campaign calls for everyone – individuals and businesses alike – to switch off their lights for one hour. In 2017, the campaign’s 10th anniversary, millions of people and thousands of landmarks around the world turned their lights off for a single hour. The WWF uses the #EarthHour hashtag (amongst others) to galvanize its followers. Elements of this successful campaign are the limited time frame which makes engagement actionable, the push to share individual actions on multiple platforms, the enticing language and countdown timers.

The Cyber Harassment Helpline is an accessible toll-free Helpline for victims and survivors of online harassment and violence in Pakistan. Digital Rights Foundation founded this helpline in 2016 in response to the increasing examples of harassment of social media users and in particular women. The helpline focuses in particular on marginalized groups in Pakistan and prefers not to communicate via social media platforms for reasons of privacy and confidentiality.