Social media and the spread of fake news during a social movement: The 2019 Anti-ELAB protests in Hong Kong

The proliferation of “fake news” has been one of the most prominent concerns among journalism and political communication scholars in recent years. In one sense, the spread of inaccurate information or ungrounded speculations in times of crises is an age-old phenomenon. Shibutani’s (1966) classic formula posited that rumor arises when anxiety combines with uncertainty. When people face an uncertain situation and become highly anxious about it, and if accurate and trustworthy information is unavailable, gossips and rumors will circulate to fulfill people’s need for orientation. Rumors or fake news may be effectively quelled by credible institutions responding quickly and properly. But it also means that the problem of rumors and fake news can be particularly intractable when levels of institutional trust are low.

Of course, in contemporary societies, fake news is also a political phenomenon in that they are often intentionally generated and spread to discredit specific groups or individuals. More importantly, the political cultures in many contemporary societies are marked by increasing political polarization, the weakening of the culture of truth-seeking, and the phenomenon of normative disinhibition, that is, political actors increasingly ignore the conventional norms of how one should speak and behave in the public arena (Roudakova, 2017; Waisbord, 2018). These factors combine to create the condition in which people who spread the fake news often remain unsanctioned, and many citizens simply opt to believe what they want to believe.

Moreover, the impact of digital media transformation cannot be ignored. It is by now a well-established argument that the Internet has generated a high-choice media environment in which media outlets become more opinionated and partisan (Bennett & Iyengar, 2008; Sunstein, 2017). Researchers continue to debate the extent to which social media users are embedded in “echo chambers,” but it seems undeniable that people are at least relatively less likely to encounter disagreeable views via social media than via conventional news media that adhere to the professional norms of objectivity and balanced reporting. Meanwhile, digital and social media lead to the collapse of social contexts and/or the boundary between public and private arenas (boyd, 2002). Privatization of public communication may contribute to the fake news phenomenon by further undermining the extent to which the communication of public information is governed by the conventional public norms of truthfulness, transparency, and accountability.

In sum, there are social-psychological, political, and communicative bases for the fake news phenomenon. It is not surprising that the huge protest movement against an extradition bill in Hong Kong (Anti-ELAB Movement hereafter) in the latter half of 2019 was also marked by widespread circulation of a wide range of rumors and fake news. There is virtually no way to trace all rumors and fake news being spread during the protest movement. But the author conducted a content analysis of two fact-checking Facebook pages (Kauyim and TrueNews) and found that the two pages published a total of 549 posts between June 1 and December 31, 2019. Among the posts, 70.3% involved the countering of a piece of rumor or fake news that had a specific target of accusation, thus suggesting that many rumors and fake news could have been used to discredit specific groups and individuals; 21.5% of the fake news or rumors concerned targeted the Anti-ELAB movement, whereas 40.3% of the fake news or rumors targeted the government or the police.

That is, there are fake news and rumors accusing both sides of the contention. For example, one piece of fake news accusing the protest movement was constituted by claims saying that there were female protesters being cheated into offering free sex services to male protesters. The claims originated around late August. Originally only an “online rumor,” it gained much media attention after an ex-government official talked about it in a radio interview. The ex-official failed to provide any evidence, but said that “the daughter of a friend’s friend” was a victim. Although most mainstream media outlets dropped the story very quickly because of a lack of substantiating evidence, references to the rumor continued to appear in some pro-government media outlets. Moreover, the rumor continued to circulate online, especially through WhatsApp, often in association with pictures of nude females taken from pornographic movies. Fact-checking agency Kauyim repeatedly pointed out the “fakeness” of the photos, but the circulation of the rumor apparently did not stop. In a survey conducted by the Chinese University of Hong Kong in mid-October ¹ 2019, 9.8% of the respondents still believed the claim “there were female protesters cheated into offering sex to male protesters” to be true, and 13.3% said “don’t know.” Belief in the rumor varied by political stance. Among movement participants, 93.8% believed the claim to be fake. Among supporters of the government, only 44.3% believed the claim to be fake.

The above instance showed how one public figure’s and a few news organizations’ willingness to break the norm of reasoned public discourse (i.e. do not spread rumors without concrete evidence and do not keep reporting on ungrounded accusations) can facilitate the politicized circulation of a rumor in the mainstream media and enhance its circulation in private-oriented social media. Nevertheless, not all rumors and fake news involved unscrupulous attempts to manipulate public opinion. Take the rumor about the police killing protesters within a metro station on August 31 as the example. In the evening of August 31, after a protest march on the Hong Kong Island, conflicts broke out between a group of protesters and pro-government citizens in a metro train. Police arrived at the Prince Edward station and tried to arrest the protesters within the station. The police’s use of violence against the protesters was broadcast via television and online outlets. After the clash, the police drove all journalists away and sealed off the metro station for 3 days, and there were discrepancies in the number of injured people sent to the hospitals given by various government departments. There were rumors that the police had beaten some protesters to death within the station and had to clear the station in order to handle the corpses.

Despite the lack of any concrete evidence about people being beaten to death in the metro station, the rumors continued to circulate. In subsequent rallies and marches, protesters would shout “beating people to death on August 31” as part of a slogan against the police. In the aforementioned survey, 52.5% of the respondents believed the statement “there were protesters died when the police entered the Prince Edward station on August 31” to be true, whereas 30.9% regarded the statement as false. Among movement participants, 75.7% believed the statement to be true. Among pro-government citizens, only 8.2% believed the statement to be true.

Nevertheless, a cursory review of online discourses showed that movement supporters were aware of the lack of evidence about deaths in the station. Yet movement supporters argued that the burden of proof is on the government and the police. To continue to circulate and mention the rumor is to give pressure to the government. Here, movement supporters were using rumors as “weapons of the weak” to challenge the legitimacy of the authority (Scott, 1985). It is part of an act to reject the definition of reality promoted by the established power. This is not to say that all pro-movement rumors and fake news can be understood likewise. But what the example suggests is the need to evaluate the function and normativity of specific rumors and pieces of fake news in relation to how they are embedded in the dynamics of political contention.

Regardless of the characteristics of the fake news or rumors concerned, the aforementioned survey findings suggested that political attitude is an important determinant of whether people would take a piece of fake news or rumor as real or not. In addition, media use also influences people’s belief in rumor and fake news. I have reported in a separate article a multivariate analysis using demographics, trust in government, movement participation, support for political factions, and media use to predict whether people hold correct cognitions about certain incidents in the movement (Lee, 2020). The analysis distinguishes among three types of media: traditional media refer to mainstream newspapers and television news, movement media refer to media platforms that are known to be providing mainly pro-movement information and messages, and messaging apps refer to applications for interpersonal communication such WhatsApp and Telegram.

The findings show that consumption of movement media positively relates to holding correct cognitions that are favorable to the movement, but it negatively relates to holding correct cognitions that are unfavorable to the movement. When correct cognitions favorable to the movement and correct cognitions unfavorable to the movement are combined into overall amount of correct cognitions, consumption of movement media no longer relates significantly to the dependent variable. In other words, consumption of movement media did not lead people to have more or less correct cognitions overall. It tended to inculcate into users’ beliefs—factually correct or not—that are favorable toward the movement.

In comparison, consumption of traditional media does not relate to any of the three dependent variables. Meanwhile, use of messaging apps relates significantly negatively to correct cognitions favorable to the movement. It also has a negative coefficient in the regression model explaining correct cognitions unfavorable to the movement, even though the coefficient fails to reach the conventional level of statistical significance. Combined together, use of messaging apps relates significantly negatively to the overall amount of correct cognitions. That is, use of message apps indeed led people to hold lesser amount of correct cognitions. Since messaging apps are often used by people to connect with their personal acquaintances, the findings are consistent with the theoretical argument that privatization of public communication is a factor contributing to the proliferation of fake news and rumors.

What are the possible consequences of belief in fake news? Obviously, beliefs in different kinds of fake news may influence people’s overall attitudes toward the movement and the government. Among movement supporters, analysis of our protest onsite survey data ² further shows that protesters who believed in rumors about government and police misconduct tended to exhibit higher levels of solidarity with the militant protesters in the movement and higher levels of support for a range of radical actions. These findings suggest that the circulation of rumors and fake news among protesters may serve the purpose of sustaining the outrage and thus the protest movement at large.

To recapitulate, the fake news phenomenon is understood as being grounded in the social-psychological needs of people in times of uncertainty, a political culture marked by polarization and normative disinhibition, and a mediascape that facilitates the fragmentation and privatization of public communication. Fake news has specific roles to play in the context of contentious politics. It can be used by political power to delegitimize a protest movement, but it can also be used by a protest movement to pressurize the political power and to sustain itself. Many issues await further examination. But one principle derived from this discussion is that the roles, consequences, and normative desirability of specific pieces of fake news and rumors need to be evaluated and examined in terms of how they are embedded in the power relationships and interactional dynamics of the protest movement concerned.