Executive Summary

Disinformation campaigns on social media are particularly detrimental to the healthy functioning of a democracy, given the increasingly central role of social media in political communication^[1]. A growing number of governments have been deploying disinformation campaigns on social media for the manipulation of public opinion and agenda-setting, two practices that hamper the proper functioning of a democracy.

In recent years, disinformation campaigns on social media have been deployed by some governments for the manipulation of public opinion and agenda-setting, two practices that hamper the proper functioning of a democracy.

Political astroturfing is defined as “a centrally coordinated disinformation campaign in which participants pretend to be ordinary citizens acting independently.”^[2] In political astroturfing campaigns, state-sponsored trolls “appear to be part of a genuine grassroots movement or sentiment” but actually are members of a top-down organization focused on intentionally spreading disinformation.^[3]

In June 2020, Twitter suspended and removed 7,340 accounts attributed to the youth wing of Turkey’s ruling Justice and Development Party (AKP) because of their organized and coordinated inauthentic activities to promote the AKP’s political agenda and target the opponents of the government.^[4] During that same period, Twitter disclosed data it had accumulated on 23,750 Chinese and 1,152 Russian state-sponsored accounts. The majority of the Turkish accounts released by Twitter were fake and compromised accounts created to disseminate political narratives favorable to the AKP and denigrate opposition parties and groups.

The existence of the troll accounts became public knowledge during the Gezi Park protests in 2013.^[5] The AKP and its trolls, backed by state and local municipal resources, eventually became the dominant group on social media compared with groups that were opposed to the AKP. The social media has increasingly been weaponized to suppress dissent and vilify opponents of different social affiliation and political persuasion.

The current report aims to shed light on various aspects of the AKP’s political astroturfing campaigns by conducting an in-depth analysis of the AKP-sponsored Twitter accounts and their tweets in Turkey.

Our analysis showed the extent and modus operandi of the AKP government’s political astroturfing campaigns through social media during elections and critical national event. The main goals of these campaigns were to dominate public discourse, set the public agenda in accordance with the best interests of the AKP administration, and vilify and discredit political discontents.

The most remarkable method used by the troll accounts in these campaigns was masquerading as different political identities to disseminate their messages and false information among various segments of Turkish society and across the political spectrum. Based on self-identifications and the content of the tweets, we grouped the AKTroll accounts into four categories: pro-Erdogan, Kemalist, Nationalist, and pro-Kurdish. These four categories represent the major political groups in Turkey as they are currently represented in the Parliament.

The troll account-creation trends of timing, subject-matter, and political affiliation led to the conclusion that reflects the AKP’s strategic priorities. A closer look at the tweets from the top trolls sheds light on the AKP government’s political astroturfing campaigns. In the most retweeted posts that we analyzed, the top troll accounts:

• disseminated false information both in favor of the AKP government and against the opposition parties and groups.
• invited their followers to tweet about pre-determined topics so that they would be trending across social media ecosystem.
• portrayed opposition parties and key figures as allies and apologetic defenders of certain groups (e.g., Kurds, Gulen Movement members, minorities), who were declared as political outlaws by the government in a rapidly-shifting political context.
• sought to justify the AKP government’s most controversial policies and discredit several legal probes launched into some political officials’ misdeeds (e.g., corruption, bribery, human rights violations).
• incited and justified violence, cruelty, and injustice against some members of opposition groups.

These methods became more visible before and during elections and critical national events that unsettled the political contours in the past decade can be described as follows: the Gezi Park protests (2013), the corruption investigations (2013), the Kobani protests (2014), and the July 2016 coup attempt.

The troll account-creation trends of timing, subject-matter, and political affiliation led to the conclusion that reflects the AKP’s strategic priorities. The first significant peak in the number of new troll accounts occurred during the Gezi Park protests in 2013.

The next sudden increase in new accounts coincided with the December 2013 graft investigations. The Kobani protests across Turkey were the manifestation of Kurdish displeasure against the government’s policy in northeastern Syria. Around the time of the July 15, 2016 coup attempt and the national rally in Istanbul’s Yenikapi district that followed, new troll accounts were created en masse. Between the July 2016 coup attempt and the 2019 local elections, scores of new troll accounts were set up.

The 2019 local elections coincided with the highest surge in the number of new troll accounts created to spread disinformation. The AKP’s defeat in the municipal elections prompted a reorganization of Internet operations.

The trends in the number of newly created troll accounts posing as members of non-AKP political parties or subscribers of different political ideologies overlap with key national events that posed daunting challenges to the AKP government’s hold on power. Kemalist accounts are deployed as part of the AKP’s underlying strategy of political astroturfing on social media to create a false impression that Kemalists genuinely support AKP government policies.

The AKTroll organization is not transparent. Although the organization used resources from the national and local governments, their organizational structure has never been officially disclosed.

Although the “code of AKTrolls” and the troll armies’ organizational structure are not fully known, the trolls disguise themselves in different sockpuppet typologies, or allegiances.

The timing and content of tweets corroborate that leading AKP officials mobilized such accounts to the advantage of the party. This phenomenon was observed countless times during the key national events elaborated above.

The trolls benefit from the convergence of government resources and media power. After the 2019 municipal elections, the financial support the AKTrolls were used to receive from local administrations considerably began to erode.

Another transformation of the troll armies could be the expansion of their operations from the domestic landscape to foreign misinformation and disinformation campaigns.

Turkey’s democratic breakdown undercuts media freedoms, while social media appear the last lines of defense. Turkey’s pivot away from democratic path has only exacerbated the precarious state of media freedom. This leaves social media as the last contested battleground where critical voices speak out.

For the silenced opposition groups and the human rights defenders in Turkey, social media has been left as the last stand against the ever-expanding political encroachment over the boundaries of individual expression. It is the last viable platform where different people from all walks of life voice their grievances about the ongoing injustices and their personal plight in the country.

Given the high Internet penetration rate in Turkey and the extensive use of social media by the tech-savvy Turkish people, the control of social media has become an imperative for the government, which deeply loathes any free room for dissent. Our findings confer credibility on the assessment that bears resemblance to a number of authoritarian governments such as Russia, China, and Iran in terms of political astroturfing campaigns in domestic realm.

Introduction

The dawn of post-truth politics across the world in the past decade has called into question the role of social media tools and other communication techniques in the demise of truth. The Western academia, intellectuals, and democrats have come to reckon that the much-cherished social media has not been as progressive as many people have imagined. This issue took an air of urgency after Russia’s perceived role in the election campaign of Donald J. Trump in 2016 that had been revealed by the U.S. intelligence agencies. In addition to the weaponization of new-age media tools in the geopolitical tug of war between global powers, the decline of truth and the pervasive contagion of disinformation across social media sphere have been thrown into the heart of political debate.

This aspect has been acknowledged by Silicon Valley giants such as Facebook and Twitter as they, after intense public and political pressure for action, finally moved to shut down the fake accounts, bots, and suspicious users who act on the payroll of national governments. Last year, Stanford University released an extensive study documenting how foreign governments sponsored disinformation campaigns on Twitter.

Stirred by political and public pressure, Twitter terminated thousands of accounts in China, Russia, Venezuela and Turkey last year, citing government links. Building on the Stanford University’s report, this study explores the political astroturfing campaigns on Twitter and how they impact the public opinion before and after critical national events that concern the entire nation in the case of Turkey. Using an empirical method, this study seeks to weave together a coherent narrative after building a correlation between the data that documents Twitter activity and the actual political affairs that form the basis of such tweets.

This report has four sections. The first section discusses how disinformation and manipulation campaigns on social media have emerged as significant threats to democracies and the notion of truth in recent years. The second section describes the historical dynamics that led to the rise of the Justice and Development Party (AKP) in Turkey, investigates the institutional transformation of the government bureaucracy, and explains Turkey’s democratic backsliding after 2013. The third section provides a more in-depth explanation of how the AKP sought to subvert the traditional media by redesigning the media landscape. This section also demonstrates AKP’s subtle and overt efforts to shape social media. The final section presents our findings on the analysis of 6,252 troll accounts and their tweets−the patterns and trends of the trolls’ disinformation activities and their modus operandi around elections and other critical political events.

Disinformation, Manipulation, and Political Astroturfing

“History stopped in 1936– after that, there was only propaganda.”^[6] As George Orwell’s famous phrase reveals, disinformation, manipulation, and propaganda are not recent phenomenon. They were building blocks of fascist, authoritarian, and totalitarian regimes well before the recent dawn of the post-truth politics. Although the perennial war over truth is as old as human history, the rise of alternative facts is mostly a consequence of the 20th century political ideologies and currents. The re-emergence of alternative facts^[7] took place after the ascent of Donald J. Trump’s to the White House in 2016. And since then, it has become an element of contemporary politics.

This new pattern has enabled the political recourse to disinformation, manipulation, and degradation of truth as useful weapons to redesign the boundaries of political and public discourse.

The term disinformation is defined as providing “false, inaccurate, or misleading information designed, presented and promoted to intentionally cause public harm or for profit.”^[8] According to this definition, three main criteria determine disinformation: (1) deception, (2) potential for harm, and (3) an intent to harm.

In a broader sense, manipulation refers to discourse used to mobilize people to act in “the best interest of the dominant group and against the best interests of dominated groups.”^[9] It also means the “communicative and interactional practice in which a manipulator exercises control over other people, usually against their will or against their best interests.”^[10] Politicians, for example, may use manipulative discourses to control people’s opinions, ideologies, and actions through the abuse of social power and domination.^[11]

Political campaigns on social media that use disinformation and manipulation might be particularly detrimental to the healthy functioning of a democracy, given the increasingly central role of social media in political communication ^[12] and the decreasing level of public trust in traditional media. Research has shown that low levels of trust in the political system and media make people more open to exposure to disinformation campaigns.^[13] As public trust in media has declined in recent years and the political tribalism has deepened across social spectrum, supporters of certain ideologies and political parties have started to rely more on social media networks rather than traditional news sources to access and share information. This trend has created “dedicated audiences” that political powers can use for “the delivery of identity solidarity” rather than “informed discourse.”^[14] The increasing use of social media has provided invaluable platforms to politicians to create identity solidarity, manipulate dedicated audiences, disseminate false information, and set the public agenda in accordance with their interests.

In recent years, disinformation campaigns on social media have been deployed by some governments for the manipulation of public opinion and agenda-setting, two practices that hamper the proper functioning of a democracy. ^[15] The level of danger caused by disinformation on social media was recognized only after Russian interference in the 2016 U.S. presidential election. Thousands of social media accounts—including those on Twitter, Facebook, and Reddit—were made to look like they belonged to ordinary U.S. citizens. These types of campaigns are called political astroturfing. It has recently become a widely-appealed disinformation methodology by authoritarian regimes.

Political astroturfing is defined as “a centrally coordinated disinformation campaign in which participants pretend to be ordinary citizens acting independently.”^[16] In political astroturfing campaigns, state-sponsored trolls “appear to be part of a genuine grassroots movement or sentiment” but actually are members of a top-down organization focused on intentionally spreading disinformation.^[17]

Regardless of whether the information is true or false, the key strategy in political astroturfing campaigns is to give the public the false impression that there is a genuine popular support for the ideas they share on social media.^[18] The potential for social media to set the news agenda is no longer a far-fetched idea, but within the realm of possibility given that journalists, politicians, and the public pay close attention to the trending topics on Twitter and other social media platforms.^[19]

The term agenda-setting has been used to refer to the power of traditional news media to create public awareness of and concern about most pressing issues.^[20] To achieve this goal, media selectively omits or overrepresents certain issues during its coverage to make the public think that those issues are more important than others. Consequently, the gatekeeping role of traditional media is an indispensable part of its agenda-setting power. But this monopoly over determining what is to be talked or not (at the national scale) is no longer taken for granted. The viral impact on social media platforms has made dissemination of information with lightening speed possible. The direct consequence of this new phenomenon is the unseating traditional outlets from their previous esteem and power by allowing outsiders (individual or collective) to orchestrate political astroturfing campaigns and set the public agenda easier than before.^[21]

Political astroturfing on social media differs from the use of automated bots to enable a topic to become trending. Astroturfing campaigns usually are waged by actual humans who are on the payroll of governments or political powers. The downside of deploying human trolls is that it requires active editorial policy and meticulous management of human resources. It is also more expensive than automated bots. But it is not entirely without benefits (according to the adherents of these campaigns). Instead of letting the algorithms run the show, human trolls are in charge in political astroturfing. The upside of cashing in on actual people is the perceived view that humans are considered more efficient for their ability to custom-tailor a campaign sometimes in minute details in order to influence the target audience. A recent study reveals that many of the political astroturfing campaigns on social media are run by different governments: Russia, North Korea, China, Venezuela, and Iran.^[22] In these campaigns, the common goal of the individual governments is to mislead the public by creating “a false impression that there is genuine grass-roots support or opposition for a particular group or policy.” These campaigns use a similar strategy of tweeting about a pre-identified topic during the same period, as trolls are briefed in advance about the content and timing of the tweets well before the campaign begins.

A common technique used in the political astroturfing campaigns on social media is sockpuppetry, which refers to political trolls’ practice of creating false identities on social media to conceal their true identity and motives^[23]. The main goal of sockpuppet accounts is posing as members of various social or political groups to serve the purposes of the political astroturfing campaigns (i.e., to spread misinformation, harass or harm the targets, set the agenda in favor of the dominant group).

In June 2020, Twitter suspended and removed 7,340 accounts attributed to the youth wing of Turkey’s ruling Justice and Development Party (AKP) because of their organized and coordinated inauthentic activities to promote the AKP’s political agenda and target the opponents of the government.^[24] During that same period, Twitter disclosed data it had accumulated on 23,750 Chinese and 1,152 Russian state-sponsored accounts. The majority of the Turkish accounts released by Twitter were fake and compromised accounts created to disseminate political narratives favorable to the AKP and denigrate opposition parties and groups.

A report by Stanford University researchers on pro-AKP accounts stated that these accounts were centrally managed for “AKP cheerleading” about certain policies and shaping public opinion during critical national events such as the 2017 constitutional referendum and Turkey’s 2019 offensive in northern Syria.^[25] The trolls, according to the methods disclosed by the report, created fabricated personalities with similar usernames. In the next step, the same accounts then created pro-AKP “retweet rings” to promote certain accounts and tweets. By doing so, this self-referential method thereby facilitated the building of an “inauthentic audience” and falsely popularized the hashtags created during a specific campaign. The HDP (Peoples’ Democratic Party) and the CHP (Republican People’s Party), the report findings demonstrate, appeared to be the chief victims/targets of such campaigns.

The AKP’s first extensive deployment of troll accounts took place during the Gezi Park protests in the summer 2013. An ethnographic study on AKTrolls^[26] indicated that the trolls had an impact on shaping and countering political discourse on social media since the protests began. The AKP’s trolling practices included (1) aiding the surveillance activities of the government, (2) undertaking the social lynching of dissidents and individuals who openly criticized the government, (3) acquiring effective and famous social media accounts to refashion them for their own purposes, (4) using automated bots to counter anti-AKP discourse, and (5) targeting foreign nationals who were critical of AKP policies.

While the Stanford University report presents only the tip of the iceberg considering the amount of troll data shared by Twitter, it is an important revelation of the AKP’s political astroturfing campaigns on Twitter. The current report aims to shed light on various aspects of the AKP’s political astroturfing campaigns by conducting a more in-depth analysis of the 6,252 government-sponsored Twitter accounts and their tweets in Turkey.

The Genesis and Rise of AKP

The 18-year of Justice and Development Party (AKP) rule in Turkey (2002-2020) has been marked by initial phases of novel democratization reforms and inexorable ebb of democracy, human rights, and freedom. During its first decade in power, the AKP mostly honored its promises by pursuing a reformist agenda and taking concrete steps to materialize Turkey’s decades-old bid for a full membership in the European Union.^[27] The country was branded by the United States as a role model in the Middle East.^[28] During the Arab Spring, the Arab streets regarded Turkey as a source of inspiration and a model to be followed.^[29] Then-Prime Minister Erdogan was among the most admired world leaders according to the Arab public opinion, a regional survey by Washington-based Brookings Institute revealed in late 2011. It did not take long for the unraveling of the much-cherished Turkish model.

The first cracks within this model began to emerge when the AKP’s commitment to the course of democracy started to be dubious and feeble in 2013. This anti-democratic shift first manifested itself during the Gezi Park protests the same year when the government moved to crush the popular discontent with a heavy-handed response. As the government used the protests as a pretext to consolidate his power base by invocation of the divisive us-and-them rhetoric, the Gezi Park crackdown became a crucial turning point for the AKP’s subsequent political trajectory. The government’s commitment to civic politics and democracy seriously began to erode. The Gezi also drove a wedge between Ankara and the West, generating the loss of goodwill from the Western supporters of Turkey.^[30]

In the post-Gezi era, the AKP administration has displayed a chilling appetite for rejecting political plurality and has shown a penchant for consolidation of power over various sectors of the government bureaucracy, the judiciary, the military, and the media. These institutions thus became the AKP’s principal means for consolidating its authority over the political, social, and economic segments of society and establishing a “competitive authoritarian” regime in Turkey.^[31]

To streamline this piecemeal transformation as smoothly as possible, the mastery of the media and public discourse has become a matter of top priority for the ruling party. As it moved farther and farther from the democratic ideals it has once championed,^[32] the government tightened its grip on traditional media outlets through a combination of methods. It intimidated some media bosses to sell their outlets to pro-AKP conglomerates to expand its sway over mainstream media landscape. This required subtle policies that involved administrative and legal measures at its disposal through TMSF (a semi-autonomous government agency responsible for managing the affairs of financially unstable private companies). The second government strategy was more blunt. The government directly seized the assets of independent media groups, appointing trustees to run the day-to-day management of outlets after seizure.^[33]

The last measure undertaken by the government had no precedents in Turkey’s modern history. Using the abortive 2016 coup as a pretext, the government shut down at least 180 media outlets, imprisoning scores of critical journalists en masse in a bid to subdue any form of dissent in media. This collective punishment and total war strategy for absolute control over media has been documented by a number of international watchdogs in detailed fashion. Yet, none of these methods, however pervasive or all-encompassing, would do little good to control social media. For this reason, the AKP administration increasingly relied on alternative tactics to regulate social media to bend the public narrative to fit its political purposes.

The government has correspondingly channeled its resources to these communication technologies in an effort to expand its influence over different platforms of social media by orchestrating trolling and political astroturfing campaigns.^[34] Political astroturfing refers to the airing of state-sponsored political campaigns on social media by using trolls that appear to be part of a genuine grassroots movement but actually are part of a centrally orchestrated top-down organization. (China’s “Fifty-Cent Army” is one of the best-known cases of politically-engineered trolling on social media.) In this study, we analyzed data from the Twitter Transparency Agency to demonstrate how the AKP government used trolls (also known as AKTrolls) before and during key political events to disseminate false information, manipulate public opinion in the country, and set the public agenda in accordance with the AKP’s political interests.

The AKP’s Growing Sway Over Social Media

The crown jewel of all achievements the AKP has obtained during its long spell in power is its mastery of the country’s media landscape. This is no small triumph. The AKP almost eradicated the space for independent media and created a pro-AKP media network through intimidation, coercion, and unlawful shutdowns.^[35] The AKP encouraged loyal businesspeople to purchase media outlets while purging critical voices in the media by arresting hundreds of journalists. The 90 percent of media openly promotes the AKP’s political agenda and serves its ideological purposes.

While the AKP’s coercive and unlawful tactics have put the most of Turkey’s traditional media outlets under the full control of the government, this is the not case for social media. The ownership and operating structure of social media platforms completely differ from traditional outlets. Given that social media is originated elsewhere (mostly in Silicon Valley in the U.S.), it deprives national governments of usual measures and channels to shut them down easily on arbitrary whims. This set of advantages made social media an attractive choice for millions of users, who seek an escape from government regulations and the gatekeeping role of traditional media editors, around the world. Its meteoric rise took place during 2009 streets protests in Iran, the Arab Spring in 2010 and 2011 across the Middle East and North Africa (MENA), and 2013 Gezi Park protests in Turkey where social media’s popularity surpassed the traditional media in a public display of defiance and anger. Naturally, the growing appeal of social media manifested itself in the political context of social upheavals and popular protests across the world during this period. It was not lost on the AKP government, which quickly came to reckon with the lure and widespread impact of this new media tool. Given the power gap between the old and new media, and the increasing popularity of social media—particularly among youth—the AKP established a pro-government troll army, or AKTrolls (i.e., pro-AKP political trolls), to project a pro-government narrative on new layers of the public discourse on social media by manipulating its very operating existence.

Troll Army: AKTrolls

The existence of the AKP’s secretive army of trolls became public knowledge during the Gezi Park protests in 2013.^[36] The opposition had also been using social media during the protests, but their goal was to organize and mobilize the masses. The AKP, however, chose to use social media—especially Twitter—for spreading pro-government propaganda. Members of the troll army were “pro-government internet users whose productive engagement with the authorities through social media networks can be seen as a form of digital surveillance, which in turn triggers restrictive consequences for citizens located in the ranks of Turkey’s opposition.”^[37] The AKP and its trolls, backed by state and local municipal resources, eventually became the dominant group on social media compared with groups that were opposed to the AKP. The democratizing effect of Twitter by allowing more people to engage in civic discussion and political debate was soon offset by the government deployment of trolls who were created for the exact reason to kill that prospect for a proper national dialogue. Gone were the days when social media platforms—Twitter in particular—were competitive forums for AKP followers and the opposition.

The government’s initial motive for creating the troll army was to counter the Gezi Park protestors’ social media campaign. It was rather a reactionary and defensive strategy against what the government perceived to be the protestors’ very effective use of social media platforms to mobilize the masses, coordinate protests in almost every city in Turkey, and engage with the outside world to win support from the international community. Witnessing the protestors’ success, the AKP understood the critical role that social media could play in advancing its own objectives.

The AKP’s second, and perhaps last, responsive and defensive engagement with social media was during and after the December 17 and 25, 2013 corruption investigations that implicated Erdogan’s family and his close associates. Since 2014, the AKP has been on the offensive. It has been in the leading position in social media sphere vis-à-vis discourse control, the manipulation of public perception, and the pressuring and lynching of opposition figures—including politicians, intellectuals, civil rights activities, and journalists.

The social media has increasingly been weaponized to suppress dissent and vilify opponents of different social affiliation and political persuasion. Although the Internet penetration rate, which corresponds to the percentage of the population that uses the Internet, is not high in Turkey compared with other countries, “the use of social media is extremely popular among Turkish Internet users.”^[38] According to a report by the Center for American Progress, the distrust toward traditional media in Turkey led the public to rely more on social media—resulting in a fivefold increase in the penetration rate from 2 percent in 2015 to 10 percent in 2018.^[39] The AKP, therefore, sees social media as an important tool to sustain popular support.

The AKP has attempted to expand its sway over social media through government oversight, and heavy financial fines against foreign companies, such as Facebook and Twitter, to bend them to serve its purposes. According to a recent legislation enacted in October 2020, “social media providers with over a million Turkish daily users must open offices or appoint a legal representative in the country. If companies do not comply with the new rules, they face major fines.”^[40] Furthermore, the opposition figures’ access to social media has frequently been blocked or limited by passing new laws and regulating Internet service providers.^[41] The AKP’s double approach to the regulation of social media means that while the opposition’s access to social media faces repeated restrictions and limitations, state agencies are able to purchase and activate “software-based blocking and surveillance” tools.^[42]

The AKP’s campaign to regulate social media landscape consists of a three-part strategy. The first one is related to the government’s frequent invocation of the criminal law articles to regulate the social media. In this respect, any political discontent can be prosecuted according to vaguely defined laws. A number of notorious laws contain provisions such as “insulting the President” that is set out by a Turkish Penal Code prohibition and that is used as a cudgel against zillions of social media users. Turkey’s infamous anti-terrorism law is another legal instrument in this equation. The law normally prohibits a person from “making propaganda for a terrorist organization”^[43] and prohibits someone from “inciting enmity or hatred.” But it offers a wide latitude to prosecutors to go after anyone who dare to criticize President Erdogan’s personality or the government policies. According to the Human Rights Watch, “Turkish courts have convicted thousands of people in the past four years simply for speaking out against the president.”^[44]

Secondly, the AKP enacted new laws to seemingly legalize social media platforms. Behind the facade of this regulation frenzy, these laws are actually designed for pressuring, intimidating, or punishing social media companies to ensure that they comply with the demands of the government. For example, the Law on the Arrangement of Internet Publication and Combating Crimes Committed Through These Publications allows the government to “impose stringent requirements on social media companies and bolster the government’s ability to censor online speech.”^[45] The law requires foreign social media companies to establish a legal representative in Turkey and to store data on local servers; prohibits such companies from advertising; and imposes fines of up to $700,000 on companies that fail to remove certain content from social media platforms.^[46] These new measures place social media companies in an unpleasant situation, mostly at the mercy and whims of the Turkish government if they fail to toe the line.

The third strategy involves the use of troll armies, or AKTrolls, against opposition figures. The intent is to discredit and defame paramount opposition figures, while also aiming to hinder the opposition’s general engagement on social media. The trolls are responsible for laying the groundwork for the AKP’s future political interventions (i.e., specific policy reform campaigns), controlling the political discourse, manipulating public opinion, directing the public agenda, creating political pressure, lynching the opposition or any target of interest, masquerading as a political ideologue or activist, and mobilizing political and social support bases.

In the following section, the data and analytic processes are used to examine the AKP-linked trolls’ tactics to whip up public sentiment for the government’s political causes, manipulate political climate, undermine civic discourse, and fine-tune the government’s narratives about critical national events and contested election results along with the findings of those analyses.

Analysis of Twitter Data

Data for this report was obtained from the Twitter Transparency Center’s collection of 7,340 accounts the center took down on June 11 in 2020 for “being linked with the youth wing of the [AKP] party and a centralized network that maintained a significant number of compromised accounts.” ^[47] It was possible to access such data because of Twitter’s stated interest in transparency and its desire “to improve public understanding of inauthentic influence campaigns.”^[48] The company notes that it routinely makes available to the public the archives of tweets and media it believes were generated and disseminate by “state-linked information operations.”^[49] The center’s dataset on Turkey included 6,252 accounts and 37,782,998 tweets.

We processed the accounts and coded a sample of the tweets to understand how the account holders identified and presented themselves on Twitter. A total of 22,374 tweets were sampled to code the political self-display of the accounts. The selection of tweets was based on the most and the least retweeted tweets from each troll accounts and each year these accounts were active on the social media site. For example, if a troll account was active for three years, we selected two tweets (i.e., the most and least retweeted) for each year, resulting in a total of six tweets for this account. We then coded the accounts according to how they identified themselves in their account display. When an account did not have a username, the contents of the user’s tweets were reviewed to determine how to code the troll for one of the seven categories used in the study. We decided to use the term sockpuppet– as it is used in the literature- to describe several categories for the reason that the AKTrolls markedly disguise themselves on Twitter by masquerading as one of these identities (as shown below) to promote the AKP’s political astroturfing campaigns. The accounts were coded in this study under seven sockpuppet typologies:

1. Pro-Erdoğan. Accounts in this category mostly used one of the following identifying descriptions in their self-displays: Reis, RTE, Erdoğan, AKP, AK, Hedef 2023, Aktim, or Soylu.
2. Kemalist. Accounts in this category were identified by the following keywords: Atatürk, Ata, Laik, CHP, Altıok, Cumhuriyet, and Kemalist.
3. Nationalist. Accounts in this category used terms such as Bahçeli, Bozkurt, Ülkü, Türk İslam Birliği, Atsız, and Tanrı Türkü Korusun.
4. Pro-Kurdish. Accounts in this category used terms such as Kürt, APO, PKK, Serok, HDP, and Demirtaş.
5. News. Accounts in this category pose as a news account without any reference to an affiliation with a political party or ideology.
6. Apolitical. Accounts in this category purportedly pretend to be fans of sports clubs, literature, or invoke sentimental messages without any political or ideological narratives.
7. Others. Accounts in this category pose as supporters of a political party other than the AKP, the CHP, or the HDP. These supporters purported their allegiance to parties such as the Good Party (İYİ Parti), the Patriotic Party (Vatan Partisi), and the Great Unity Party (Büyük Birlik Partisi). Accounts in this category marketized themselves also as political reformists who lobbied for rescinding retirement age.

Table 1 shows the frequency distribution of troll armies by the sockpuppet category they are coded in the dataset.

Table 1. Breakdown of AKTrolls

The analyses presented in this report focus on four main areas where troll armies operated to promote the government’s political astroturfing campaigns.

Focus area 1

Analysis of the first focus area explores how the troll accounts reacted to critical incidents—other than the elections—that have been important to the AKP and Erdogan. Such incidents of national importance included protests (e.g., the Gezi Park protests in 2013; the Kobani protests by Kurds in 2014), the December 2013 graft probes, the PKK attacks, the July 2016 coup attempt, and so on.

Focus area 2

This area explores how the trolls navigated before, during and after elections in Turkey, including local and general parliamentary elections, referendums, and presidential elections.

Focus area 3

This focus area lays out an empirical comparison between pro-Erdogan troll accounts and Kemalist troll accounts in terms of Twitter activities (e.g., account creations or most frequent hashtags promoted).

Focus area 4

This area explores how most active trolls super-spread disinformation and factually-flawed information. .

The findings from each focus-area analysis are presented below.

Analysis 1: How had the AKTrolls’ account-creations trended during critical national events?

The troll account-creation trends of timing, subject-matter, and political affiliation led to the conclusion that reflects the AKP’s strategic priorities. These priorities include, but are not limited to, the following:

• Creating factions and sowing seeds of divisions among the two chief allies of the Nation’s Alliance (CHP and IYI Party). Trying to manipulate both parties’ constituencies by feeding them with distressing and demoralizing news about the political fortunes of their respective parties.
•  Portraying the leader of the CHP as a HDP supporter to alienate the CHP constituency from their leader.
• Manipulating ongoing political debates about national matters.
• Holding grab of the national discourse about significant events that define the contours of political landscape. Those pivotal events include the 2013 Gezi Park protests, the December 2013 graft scandal, the 2014 Kobani protests, the military operations against the urban wings of the PKK across southeastern Turkey in 2015 and 2016, the CHP party congress in 2016, the arrests of the HDP co-chairs (Selahattin Demirtaş and Figen Yüksekdağ), mass imprisonment of the HDP members.

The number of new accounts created each month from January 2010 to October 2019 and the precipitating critical national event for each spike in new account creations are shown in Figure 1.

Figure 1. Number of new troll accounts created per month and the corresponding precipitating critical events, from January 2010 to October 2019.

The exposition of the trends gleaned from the spikes in new account (creations) shown in Figure 1 is as follows:

• The first significant peak in the number of troll accounts occurred during the Gezi Park protests in 2013. AKP government had formed a team of social-media volunteers “to promote [the] ruling party’s perspective in a sphere dominated by protesters” and, “when the opposing camp spreads disinformation about the party, . . . correct them with valid information, always using positivelanguage.”^[50]
• The next sudden increase in new accounts coincided withthe December 2013 graft investigations that implicated Erdogan’s inner circle, several AKP ministers and their close relatives, and businesspeople. Soon after the probes burst open into the public view, the troll accounts were mobilized to win over the public narrative in a bid to discredit and vilify the police chiefs and prosecutors, who oversaw the initial proceedings, in a concerted disinformation campaign. This orchestrated campaign found its clear expression when then-Prime Minister Erdogan’s own daughter lent her endorsement to the deployment of trolls. She even pressed for their further deployment as an army to sway the public opinion against the investigations.^[51]
• In 2015, the AKP founded the New Turkey Digital Office (YDO) to monitor the social-media activity of rival political parties and manage the digital operations of election campaigns that same year.^[52]However, this did not generate any spike in new account creations. The YDO recruited most of its trolls from the AKP youth branches, placing them on the monthly payroll. To avoid legal repercussions, payments were delivered through private companies that mad business with the AKP-run municipalities.^[53] In a public display of political backing, President Erdogan even graced the wedding of a chief troll (with the nickname “Sağlam İrade” — Strong Will) in 2014.^[54]
• The Kobani protests across Turkey were the manifestation of Kurdish displeasure against the government’s policy in northeastern Syria. In October 2014, thousands of Kurdish protestors took to the streets across the country and more than 40 people were killed in the clashes between rival groups. During the Kobani protests, the number of new troll accounts spiked.
• After peace talks between the PKK and the AKP government collapsed, the government began military operations in several cities (also known as Hendek operations) that led to the deaths of hundreds of civilians in predominantly Kurdish areas between July 2015 and December 2016. The number of new troll accounts correspondingly skyrocketed during this period.
• Around the time of the July 15, 2016 coup attempt and the national rally in Istanbul’s Yenikapi district that followed, new troll accounts were created en masse.
• Between the July 2016 coup attempt and the 2019 local elections, scores of new troll accounts were set up. This pattern emerged whenever an event of vital importance (e.g., the forced resignation of the mayors of Istanbul and Ankara) took place. To make damage control in advance and shape the public opinion against any negative spillover of such key events, those accounts played a moderating role in social media sphere by projecting positive aspects of each incident.

Analysis 2: How had AKTrolls’ account creations trended around the elections between 2010 and 2019?

Based on the findings, the timing with regards to the troll account creations trend led to a set of conclusions. These reflect the AKP’s strategic priority of influencing voter behavior at the ballot box. The precipitating events, which are exhibited on the graph in Figure 2, include, but are not limited to, the following:

• The first elections held after the December 2013 corruption investigations were municipal polls in March 2014 and the presidential election in August that same year. In the electoral context, the number of new troll accounts increased significantly.
• In January 2016, the CHP organized its general party convention. That same month, the number of AKTrolls masquerading as Kemalists surged dramatically.
• In May 2016, Erdogan forced then-Prime Minister Ahmet Davutoglu to resign both from the post of the prime ministry and the AKP chairmanship. Erdogan’s close ally, Binali Yildirim, was appointed as Davutoglu’s successor in the AKP congress. During this internal power struggle and political coup d’état against Davutoglu, a peak in the number of new troll accounts was observed.
• The 2019 local elections coincided with the highest surge in the number of new troll accounts created to spread disinformation. The AKP suffered a stinging rebuke in the polls when it simultaneously lost Istanbul and Ankara in 2019. These were the two cities from where most of the funding for the troll accounts originated. This electoral debacle may explain the corresponding decline in new troll accounts in the following months after the conclusion of the polls.
• The AKP’s defeat in the municipal elections prompted a reorganization of Internet operations. While the reorganization brought takedown of some troll accounts, Erdogan’s new media Czar reportedly laid out a new strategy, direction, and a renewed purpose for the disgruntled troll army.^[55]

Figure 2. Number of new troll accounts created per month that corresponded to elections, January 2010 through October 2019.

Analysis 3: How do the monthly account-creations vary between Pro-Erdogan and ‘Kemalist’ AKTrolls during critical events?

Figure 3 lays out a comparison between the monthly trends in the creation of Pro-Erdogan troll accounts and the Kemalist troll accounts.

Figure 3. Number of new Pro-Erdogan and Kemalist AKTroll troll accounts created per month and corresponding critical events, from January 2010 to October 2019.

The trends in the number of newly created troll accounts posing as members of non-AKP political parties or subscribers of different political ideologies overlap with key national events that posed daunting challenges to the AKP government’s hold on power. Spikes in the number of these new troll accounts were, in some cases, quite high. In May 2016, for instance, the creation of troll accounts posing as genuine Kemalists went up when Erdogan cut Davutoglu loose and picked his point man, Yildirim, as new prime minister. To contain a potential fallout from such a critical change, the social media team steered a very subtle campaign to uplift the mood of the bewildered party base. The campaign also sought to thwart any potential internal insurgency from the AKP rank-and-file and former bigwigs.

Analysis 4: How do pro-Erdogan troll accounts compare with Kemalist troll accounts in their activities on Twitter?

Activity volume (i.e., retweets, frequently used words, and hashtags) of pro-Erdogan accounts drastically differs from Kemalist accounts. As shown in Table 2, the average number of retweets by pro-Erdogan accounts was 7,121, while that average for Kemalist accounts was 11,681—meaning that Kemalist accounts had about 61 percent more retweets than pro-Erdogan accounts.

Table 2. Average Number of Retweets by Account Category

Pro-Erdogan and Kemalist AKTroll accounts have been deployed to impersonate different political parties, affiliations, ideologies, and ethnicities with one or both of the following objectives in mind:

• Kemalist accounts are deployed as part of the AKP’s underlying strategy of political astroturfing on social media to create a natural impression that Kemalists genuinely support AKP government policies.
• The Kemalist accounts portray themselves as internal opponents of the CHP agenda and policies.

For example, the name Recep Tayyip frequently appeared in the word-cloud tweets of troll accounts who self-display as supporters of Erdogan (Figure 4).

Figure 4. Word cloud from tweets of Pro-Erdogan AKTroll accounts.

The word cloud from Kemalist AKTrolls in Figure 5 reveal that AKP and AKP and Evet were salient, while Mustafa Kemal is less recurrent. This was so even though these trolls displayed themselves as supporters of Mustafa Kemal (Ataturk), the founding father of the Republic of Turkey.

Figure 5. Word cloud from tweets of Kemalist AKTroll accounts.

Similarly, the hashtag clouds in Figures 6 and 7 show that troll accounts manipulate trending hashtags differently in accordance with their self-displays. For example, pro-Erdogan accounts heavily engage with the hashtags that AKP supporters initiated, while Kemalist accounts engage mostly with #Hayir (No), which the government’s political opponents usually supported before 2017 referendum. To give a context, it must be noted that this #Yes and #No campaigns took place before and during April 2017 constitutional referendum that later set the stage for a transition to the executive presidential system in Turkey.

Figure 6. Hashtag cloud from tweets/retweets of Pro-Erdogan AKTrolls.

Turkey’s opposition and the government reserved enormous resources to promote their campaigns before the 2017 referendum.

Figure 7. Hashtag cloud from tweets/retweets of Kemalist AKTrolls

The political opposition rooted for the defeat of the proposed shift to presidential system with #Hayir (No) campaign, while the government supporters wooed undecided voters with promoting #Evet (Yes) hashtags on Twitter.

Analysis 5: How had the most active AKTroll accounts operated during elections and other key events in Turkey?

The troll activities intertwined with the prevailing political narratives the AKP leadership had adopted against the political opposition. The following events, therefore, warrant further analysis:

• The aftermath of June 2015 general elections
• The CHP’s General Party Convention in January 2016
• The targeting of CHP leader Kemal Kilicdaroglu and the divisions it generated (i.e., the Muharrem Ince factor; Ince has been a rival of Kilicdaroglu within the CHP)
• The AKP Party Convention in May 2016
• The aftermath of the July 2016 coup attempt
• The police investigations in November 2016 that targeted HDP leaders
• The Halkbank trial at the United States District Court for the Southern District of New York in October 2019
• The snap elections in November 2015
• The presidential elections in June 2018
• The Pastor Andrew Brunson crisis between the United States and Turkey in October 2018

We qualitatively analyzed the tweets posted by top troll accounts during three months before elections and during three months after the key events in Turkey since 2013. The accounts that had the most retweets within any given period were considered to be the top trolls. These trolls had large numbers of followers, ranging from 78,000 to 450,000, something that enabled them to reach vast audiences quickly and efficiently.

The following section discusses two aspects of the political astroturfing campaigns by trolls before elections that occurred between 2013 and 2020 and after key national events that occurred during the same period:

1. The capacity of top-10 troll accounts to produce content for these campaigns was determined by the number of tweets, the frequency of the tweets that were retweeted more than 10 times during elections and 100 times during key events, and by the average number of tweets posted by troll accounts per day during the astroturfing campaigns. The top-10 troll accounts were selected according to the number of retweets. The top accounts with the most retweets in any given period were included in the analysis.

2. How the modus operandi for the astroturfing campaigns of the top-10 AKTrolls evolved during the shifting political context was examined.

The political astroturfing campaigns before the elections and critical events are discussed separately. The findings of the descriptive analysis of the tweets of these accounts and the thematic analyses of their tweets are presented.

Tweets of the top-10 troll accounts before elections (2014-2019)

Table 3 shows the number of tweets posted by the top 10 troll accounts during the three months before elections between 2014 and 2019, the number of tweets retweeted more than 10 times, the number of tweets retweeted more than 100 times, and the average number of tweets per day during the same period.

Table 3. Number of Tweets and Retweets by the top-10 troll accounts during the three months before each election (2014-2019)

For example, before the June 2015 general election, the top trolls posted 45,238 tweets during three-month period. Within this time frame, these accounts, on average, posted 50.26 tweets per day per person. Out of the 45,238 tweets, 3,520 were retweeted more than 10 times, and 835 were retweeted more than 100 times. These numbers indicated the capacity of these top trolls to tweet and retweet a large amount of content that could affect public opinion and voters’ decisions during the polls.

A closer look at the tweets from these top Twitter users sheds light on the AKP government’s political astroturfing campaigns before key votes. In the most retweeted posts that we analyzed, the top troll accounts:

• disseminated false information both in favor of the AKP government and against the opposition parties and groups.
• invited their followers to tweet about pre-determined topics so that they would be trending across social media ecosystem.
• portrayed opposition parties and key figures as allies and apologetic defenders of certain groups (e.g., Kurds, Gulen Movement members, minorities), who were declared as political outlaws by the government in a rapidly-shifting political context.
• sought to justify the AKP government’s most controversial policies and discredit several legal probes launched into some political officials’ misdeeds (e.g., corruption, bribery, human rights violations).
• incited and justified violence, cruelty, and injustice against some members of opposition groups.

Disseminating false information

One of the underlying functions of the troll accounts was to disseminate false information in order to advance the government’s agenda and to publicly undermine members of different political creed and social affiliation. For example, @berberoglu79, a top troll account, targeted the mayoral candidate of the Nation’s Alliance for the city of Ankara, Mansur Yavas, before the March 2019 municipal elections. The troll tweeted:

@berberoglu79

Enough is enough. Mansur Yavas deeply insulted religious people by saying “Those who are most cruel towards animals are the religious people.”

In the original statement, which is available on the DailyMotion website, Yavas argued that everything that God created as equal in nature, according to the religion of Islam. He was, however, compelled to add that some seemingly religious people’s actions against animals belie that sacred notion.^[56] The troll accounts’ tweet (shown above) appears to have distorted Yavas’ remarks to convey a false impression that the Nation’s Alliance’s chief candidate was against religion and the pious people.

Such efforts span a wider spectrum of public debate on issues ranging from politics to religion, from the era-defining coup attempt to other key socio-political events that concerned the nation.

In this respect, @F_1453, another key troll account, retrospectively sought to portray the leadership of the nationalist opposition party (IYI Party) as the initial supporters of the abortive putsch on July 15, 2016, despite that it was the other way around. The leadership, who later found the IYI Party, did not waste any time to condemn the unfolding coup attempt on the day of July 15 and have repeatedly registered their disapproval of any military intent that may seek political change via non-democratic channels since then. Still, the troll went on to tweet as such:

@F_1453

On July 15 coup attempt night, while Recep Tayyip Erdoğan was on his way from Marmaris to Istanbul, IYI party members who are slandering him now were watching the coup live on TV in their holes where they were hiding.

Who supported whom during the ill-fated putsch has remained an abiding source of political controversy, and the government used this as a cudgel to repeatedly discredit and target opposition members.

 Portraying opposition parties as allies of political outcasts

The troll accounts frequently portrayed the government’s political discontents as allies of other groups, such as the Kurds and the Gulen Movement which have publicly been demonized by authorities. One of the top-10 troll accounts, isa_sahintrk, launched a smearing campaign against Ekrem Imamoglu, the joint candidate of the Nation’s Alliance for Istanbul’s mayoral post in the 2019 municipal elections. The troll tweeted:

@isa_sahintrk

You, the joint candidate of PKK, CHP and IP! You wouldn’t tell 40 lies if you had a gram of dignity, a bit of honor, a bit of bravery, a bit of manhood.

This is only one piece of countless tweets that sought to portray Imamoglu, who later won the bitter contest for Turkey’s largest city, as supporter of terrorist groups. In a concerted effort, the troll accounts correspondingly presented him as the joint candidate of a makeshift alliance of terrorist outfits that include the main opposition CHP, IYI Party, Gulen Movement, Zionists, the PKK, and many other unrelated groups.

In the same vein, another troll, @90Savas90, blamed the CHP for being in cahoots with terrorist organizations and launched a hashtag #CHPyeKırmızıKart (#RedCardToCHP). The troll tweeted as follows:

@90Savas90

It is high time to show the red card to CHP whose politics is full of dirty partnerships and based on lies, who acts with treasons, and who is a terrorist lover.

Taken together, this tweet reflects deep-seated intent and underlying efforts of the AKP apparatus to associate the opposition CHP with terror groups in public view. These two tweets, a small representative of similar tweets, illustrate that point clearly.

Before the August 2014 presidential election, which coincided with the first anniversary of the Gezi Park protests, the AKTrolls presented the Gezi Park protests as a conspiracy against the AKP government. One of them, @durduralamaz, tweeted:

@durduralamaz

There were neither “gezi” nor “dictator” [criticisms against the government] while the interest rates were 150%, Turkey was begging for money from the IMF, and martyrdom news were coming each day. Don’t you still understand “gezi-minded”? [a derogatory term meaning “feeble-minded,” which is used by AKP supporters to insult Gezi Park protesters].

This and other tweets projected the Gezi as a concerted endeavor to unseat the government via nationwide street protests, while the state of national affairs was faring better than ever before.

Counterarguments against criticisms of the AKP policies

The trolls also were busy with producing counterarguments against public criticism of the AKP government’s policy failures. When the government’s image was irredeemably tarnished by the outbreak of corruption allegations in December 2013, the trolls were mobilized to downplay the investigation, to discredit prosecutors and police officers, and to vilify the opposition. A similar kind of textbook was shrewdly deployed to deviate public attention from the gross human rights violations that took place under the watch of the AKP officials over the past several years.

In the wake of the (ill-fated) December 2013 graft probes, Erdogan’s party regarded the legal proceedings as a judicial coup “orchestrated” by the Gulen Movement to unseat him from the power. To nurture this narrative, the troll accounts scrambled to discredit the Movement through a number of campaigns on social media. In one hashtag campaign, the top troll accounts invited their followers to support #CiamatınİçkiSofrasındaKuran, which claimed that during an event organized by the Gulen Movement, alcohol was served to participants while the Quran was being recited. The hashtag intentionally sought to conjure up an impression that the event hosts (affiliated with the Movement) disregarded the Holy Quran revered by Muslims as the revelation of God. The hashtag involves a wordplay in a subtle reference to the AKP allegation that claims the Movement worked in the same league with the U.S. intelligence agency, C.I.A., to undermine the AKP government.

Another hashtag promoted by the top troll accounts during the same period was #MontajÇetesiİhanetŞebekesi. In this campaign, the trolls sought to project the corruption tapes that leaked to the Internet as fake and “fabricated” by the members of the Gulen Movement.

In one of the tweets, the troll account, @twitt3rTURKIYE, supported the idea that the corruption tapes were assembled to disseminate false information and tampered evidence. The troll tweeted:

@twitt3rTURKIYE

A fast report by the U.S. on the [corruption] recordings: “It is totally montage. The voices were cut and past” #WeDontBuyTapePolitics

In reflection of the general AKP view, the tweet alleged that the audiotapes were products of montage rather than authentic recorded tapes. The tapes featured audio recordings of Erdogan’s murky dealings with several businessmen and with his son on legally-dubious matters, including bribery and evidence disposal.

A similar troll campaign was in place after the Soma coal-mine disaster on May 13, 2014. Some was one of the deadliest natural disasters in Turkey’s history,^[57] and the death toll (more than 300) quickly became a national scandal for the ruling AKP government. The party leaders wasted no time to win back the control of the public narrative by minimizing its role in the entire story. Correspondingly, the trolls were mobilized to attack the mainstream media for its critical coverage of the disaster, suggesting that the government actions led to the high number of causalities. The disaster took place at a time when Erdogan kicked off his campaign for the presidential election in August 2014. In a particular episode of the Twitter campaign, the top-10 trolls sought to whitewash Taner Yildiz, who at that time was the minister of energy and natural resources. The troll tweeted:

Dogan media is a provocateur and disgrace for the media. Why don’t they broadcast about the relentless efforts of Minister Taner Yildiz who have not slept for four days [since the disaster happened]?

As the government faced more and more challenges in the intervening years, the scope and scale of its political astroturfing campaigns have accordingly evolved. In each case, the troll accounts amassed considerable know-how and experience to handle the public discontent, the growing public disillusionment about a flagging economy, and other social woes that have afflicted millions.

During the three-month period before the June 2018 general elections, the main goal of the top troll accounts was to ward off the mounting public criticism against the AKP’s poor handling of a deteriorating economy. To this end, one of the trolls, @ErtugrulCagman_ tweeted:

@ErtugrulCagman_

As long as we are united, it doesn’t matter even if the US Dollar rises to 10 Turkish Liras, Euro rises to 15 Turkish Liras, and the oil price rises to 30 Turkish Liras.

But if we are divided, it doesn’t matter even if US Dollar declines to 1 Turkish Lira, Euro drops to 1 Turkish Lira, and the oil price drops to 50 Kurus [half a Turkish lira].

The oil was free in Libya, it was cheap in Syria. But there is no more Syria or Libya.

We will overcome this inshaallah (God willing)!

#YouWillNotSucceed

The Turkish lira’s free-fall against the U.S. dollar in the past years generated additional burdens for the producing Turkish business community that relied on American dollar for international transactions and for imports of goods. When the government came under increasing public criticism, the trolls scrambled to play down the currency crisis by emphasizing the national unity for maintaining Turkey’s sovereignty and self-reliance.

Inviting others to tweet on the predetermined agenda

The top troll accounts also invited their followers to tweet about a predetermined topic during the same period (i.e., the three months before the June 2018 general elections) to ensure that they could force a trending topic and shape the public agenda. Before the April 2017 constitutional referendum, the troll accounts were active in solidifying support for the AKP government’s agenda and targeting opposition parties. With the referendum 40 days away, ErtugrulCagman_ tweeted:

@ErtugrulCagman_

YES #CountDownForBigTurkey started. We will relentlessly continue to work for 40 days on the way that we know as the right one inshaallah (God willing).

Similarly, another troll, @twitt3rTURKIYE, invited his followers to support two hashtags before the August 2014 presidential election: one in favor of Erdogan and one targeting the CHP:

@twitt3rTURKIYE

TurkeyIsBetter WithErdogan: Development, Growth, Democracy

ShameForTurkey CHP: Massacres, Coup, Restrictions

The tweet sought to conjure up an association between CHP and the pre-AKP republican history marked by occasional coups, political upheavals, economic depravity, and setbacks in personal liberties. This wholesale portrait, no matter what its relation to the reality, was persistently nurtured by the trolls to undercut the CHP’s image in the public view.

Sometimes, this vilification of the opposition takes the form of openly endorsing violence against its members. This was the case when CHP chairman Kemal Kilicdaroglu was physically attacked by an AKP supporter during a funeral of a fallen soldier in 2019. YUCELKAR25, a leading troll account, heaped a lavish praise on Osman Sarıgün, the man who punched the CHP leader at the funeral.

@YUCELKAR25

Our uncle Osman Sarıgün has been released on bail. Congratulations 🙂

His arrest became a rallying cry; trolls and AKP supporters rooted for his release from police detention. The trolls loudly celebrated his release from custody.

Tweets of the top-10 trolls after critical events (2013-2020)

In similar to election periods, the capacity of the top 10 troll accounts to produce content within the span of three months following critical events was considerably high. Table 4 shows the number of tweets posted by the top troll accounts between 2013 and 2020. It also exhibits the number of tweets retweeted more than 10 times, the number of tweets retweeted more than 100 times, and the average number of tweets per day, during the same period described above.

Table 4. Number of Tweets and Retweets by the Top-10 troll accounts During the Three Months Following Critical Events (2013-2020).

During and after critical events, a remarkable volume of content was produced by the top troll accounts. After the Kobani protests, for example, the troll accounts posted 38,561 tweets during the three months following the protests, with an average of 42.84 tweets per day. Of those tweets, 2,539 were retweeted more than 10 times, and 330 were retweeted more than 100 times.

The content analysis of the tweets listed in Table 4 shows that the political astroturfing methods used in the aftermath of critical events were similar to those posted before the elections listed in Table 3.

Portraying the Gezi Park protests as a coup attempt

The AKP’s astroturfing campaign during the three months following the Gezi Park protests sought to portray the protests as a coup attempt to topple the government from power. In the context of these efforts on social media, the troll accounts draw up a resemblance between the Gezi Park protests and the military coup in Egypt that unseated President Mohammed Morsi and his government in 2013 summer.

Below are some examples of the frequently retweeted posts that both refer to the Gezi Park protests and the 2013 Egyptian coup:

@twitt3rTURKIYE

Those who were complaining yesterday about the TOMAs [anti-riot water cannon vehicles used by the police against protesters in Gezi] are greeting the military tanks in Egypt today… What a democracy!

@twitt3rTURKIYE

This [Gezi] was a real coup attempt. Those who provoked should be punished. #ArrestTaksimCoupPlatformMembers

@twitt3rTURKIYE

Did you [hear] what they did to Mursi, the coup plotted against the new elected President? They planned the same thing in the Gezi Protests, but it didn’t work. #NoToCoupInEgypt

@twitt3rTURKIYE

Frankly speaking! If our Prime Minister [Erdogan] did not stay strong against the Gezi Protests and millions did not go to the squares [to protect him], they [the protestors] would say call the army to “Mission” [Coup] #NoToCoupInEgypt

Before the military intervention, there were weeks-long street protests and violent clashes between rival political groups across Egypt. That lurked largely in the minds of the trolls who relentlessly beat the drumbeat of a coup narrative regarding Gezi protests and questioned the Gezi protests’ commitment to civilian democracy in Turkey.

Labeling the December 2013 Graft Probes as Treason

The political astroturfing campaign of the top AKP’s Trolls during the three months following the December 2013 graft probes sought to discredit the allegations that included tender fraud, embezzlement, bribery, official misconduct, abuse of power, and the obstruction of justice. This goal was best expressed by painting the investigations as a treason against and betrayal of Erdogan and the AKP government. Furthermore, the political astroturfing campaign was designed to persuade the public that Western countries and global powers orchestrated the corruption investigations through local proxies, since they wanted to thwart the rapid progress and economic development of Turkey.

They want to obstruct our development and bring Turkey on its knees while our country continues on the path towards full independence and becoming a global power.

Wake Up!

#CorruptionIsPretextTreasonIsReal

The top troll accounts echoed the idea that the leaked tape recordings that went viral on social media after the investigations were fabricated or montaged by the police. The tapes contained conversations among the suspects in the investigations and indicated how the corrupt transactions were completed (e.g., money transfers, sharing the bribery spoils among the suspects).

One of the most popular hashtags used by the troll accounts during this period was #IfTheCorruptionWereReal (#EgerYolsuzlukOlsaydi). According to the argument espoused and relentlessly spread by the trolls across social media was predicated on a hypothetical reasoning. One example of these tweets is the following:

#IfTheCorruptionWereReal, would the resources used in social services and supports have increased by 15 times [during the AKP rule]?

It mused that if the corruption allegations were real, then there would not have been any economic and social development in Turkey, which witnessed a series of remarkable accomplishments in transportation, communication, and education.

Blaming the opposition for the Kobani protests

In the aftermath of the October 2014 Kobani protests, the top- troll accounts were once again mobilized en masse to justify, endorse, and advance the AKP government’s heavy-handed response to crush the nationwide protests across the country.

Demirtaş called people to streets, CHP supported him. As a result, 23 people died. Now, they say “there is no place for violence.” The damage is already done, you murderers! [durduralamaz]

The chief objective of these Twitter campaigns was to shift the blame on the opposition parties, mostly the HDP and the CHP, allegedly for coordinating the protests, fomenting social unrest and inciting hatred.

Inciting hatred and violence against the Gulen Movement

Within the span of the three months following the July 2016 coup attempt, the top troll accounts waged a sustained campaign to promote the government’s official coup narrative and to displace alternative accounts of the abortive putsch from public debate by dominating social media. Even before any legal investigation took off the ground in the aftermath of the insurrection, the government did not waste any time to squarely place the blame on the Gulen Movement for the coup plot. Through these campaigns, the troll accounts incited hatred and justified violence against the movement sympathizers.^[58]

@YUCELKAR25

Because those who plotted the 17/25 coup attempt [the corruption probes] were not executed, they now attempted to coup with their soldiers.

Their characters are like the coyotes.

The solution is their execution.

These accounts frequently invoked the term FETO (a derogatory term that the AKP government uses in reference to the Movement) to make it a household concept since the post-coup Twitter campaign had been launched to normalize the use of the term in the public discourse.

 Blaming the opposition for the coup attempt

The top troll accounts’ Twitter campaigns during the three months following the coup attempt portrayed the opposition parties as ardent supporters of the coup attempt. For them, questioning the government’s official narrative about the botched coup amounts to treason against the country. One of the top troll accounts tweeted the hashtag #WeAreAtTaksimSquare (#Taksimdeyiz). The following tweet was posted by a Kemalist troll account, insinuating that the CHP would indeed have celebrated, if the coup attempt had been successful:

If the coup had been successful, the CHP supporters would have organized a victory rally now. #WeAreAtTaksimSquare [#Taksimdeyiz]

The troll accounts also launched a smear campaign against Meral Aksener, the leader of the IYI Party. They circulated an unproven claim that Aksener would be named prime minister, if the coup were successful. The tweet extracted Aksener’s remarks out of its context and tried to associate her with the coup. The tweet read:

@berberoglu79

Meral Aksener who said “I will march towards the palace with 25,000 people and become the Prime Minister in 6 months” is the biggest stooge of the parallel dogs.

Separately, the top trolls also kept blaming the United States, specifically CIA, for supporting and organizing the coup attempt, without bothering to offer any evidence to back up their charges.

This tactic was deployed on many occasions against prominent figures of different political persuasion in Turkey in a bid to discredit them in the public eye. Although more than four and half years have passed since the ill-fated coup attempt, the trolls still retrospectively mine who said what during and after the coup to use their remarks against them in the context of these political astroturfing campaigns.

Discussion

Misinformation and disinformation are critical components of any authoritarian regime. Regimes such as Russia, China, Iran, and Turkey use any available media tool to manipulate the domestic public opinion and divert public attention from pressing economic problems and other social woes. Accordingly, social media has become an important tool both at national and international level. The key points derived from this study are outlined as follows:

• The troll accounts are organized and paid by the AKP government to advance the party’s political objectives, agenda, and public relations strategies. This payment takes different forms. Either the AKP-run municipalities or its friendly businesses financially assist such trolling activities.
• The troll accounts’ organization is not transparent. Although the organization used resources from the national and local governments, their organizational structure has never been officially disclosed. Despite mounting public calls for more information about the workings of trolls, the government so far has not revealed any data.
• Although the “code of AKTrolls” and the AKTrolls’ organizational structure are not fully known, the trolls disguise themselves in different sockpuppet typologies, or allegiances. The way how the troll accounts steer their disinformation campaigns correspond with the definition of the term political astroturfing. First of all, the troll accounts have masked themselves as genuine supporters of a certain ideology or political affiliation. In so doing, they pose as pro-Kemalist, pro-Kurdish, or pro-Nationalist in different circumstances. Second, these disinformation campaigns endorse the AKP’s political priorities and set the public agenda accordingly, especially during election periods and after critical national events (2013-2020). Third, the available findings bolster the view that most of the campaigns seem to have been centrally coordinated, although trolls act as if they were independent and self-motivated individuals.
• Throughout political astroturfing campaigns, these troll accounts had engaged in a variety of activities to achieve their goals. In this respect, these troll accounts (1) disseminate false information against opposition parties/groups, and manipulate the public in favor of the AKP government, (2) invite others to promote centrally organized hashtag campaigns to advance the AKP agenda, (3) demonize opposition parties and groups, (4) produce counterarguments against the public criticism of AKP policy failures, (5) incite hatred and justify violence against dissidents.
• The timing and content of tweets corroborate that leading AKP officials mobilized such accounts to the AKP’s advantage.This phenomenon was observed countless times during the key national events elaborated above.
• The trolls moved from a defensive mindset to a full offensive mode after Gezi Park protests and December 2013 graft probes. It is reasonable and fair to argue that the AKTrolls have nursed their skills and know-how far better by each crisis and challenge since the outbreak of the Gezi Park protests in 2013 when the government was caught off guard against mass social media campaigns of the protesters.
• The AKP’s troll armies have several advantages. First, they are supported financially by the AKP government or AKP-affiliated businesses. Second, the trolls’ strength and intensity coincided with the election victories of the AKP and the transformation of power within the AKP. Third, internal support from the government has enabled these trolls to be proactive. Another factor is the fact while the government critics easily face prosecution for their critical tweets, trolls enjoy an unwritten policy of immunity from prosecution. They are at liberty to tweet whatever they want, without any fear for any legal backlash. This sets most of the Twitter users on an unequal footing. Last, but not least important, is the broad set of privileges accrued to the AKP’s troll army in terms of insider knowledge that is even spared from most of the journalists.
• The troll accounts benefit from the convergence of government resources and media power. This convergence enables the these trolls to circulate the same content on a vast scale, potentially allowing these AKP-government mouthpieces to act as force-multipliers, though this study did not measure that effect.
• The AKP mayors’ support for the troll armies has been diminished. After the 2019 municipal elections, the financial support the trolls were used to receive from local administrations considerably began to erode. This was on vivid display in the case of Istanbul and Ankara where a cluster of troll activities/accounts was concentrated. After the AKP lost both municipalities, the trolls lost their largest source of income. However, as the AKP fashions new ways to solidify its waning power in Turkey, a transformation of the trolls into an official entity within the government emerges as a serious prospect. If Ankara follows in the footsteps of Russia and China, the trolls can become part of the government’s communication directorate.
• Another transformation of these troll accounts could be the expansion of their operations from the domestic landscape to foreign misinformation and disinformation campaigns.The use of Twitter trolls by authoritarian governments is a relatively new practice that continues to grow in popularity. The way how Russia, China, and Turkey have formed their troll armies and operate on Twitter have striking similarities. It is unknown how much of these government-backed trolling activities relate to Ankara’s newfound alignment with Russia is a matter of mere speculation and scholarly controversy. The extent of Turkey’s international Twitter campaigns/influence has not been thoroughly and scholarly examined. (This study focused on the domestic aspects of political astroturfing campaigns within the boundaries of Turkey.)
• AKP officials and Erdogan have been very keen on asserting full control over the virtual and print media. The AKP official’s political capital and influence largely center around the party’s considerable ability to shape the political agenda and to fashion the public narrative through an all-encompassing control of the media in Turkey. Print media and television channels have fallen under enormous sway of the government through the seizure of outlets, the purchase of mainstream media channels by pliant businessmen, and unmasked political/judicial pressure. For all its penchant for total control of various branches of media, the government has failed to fully subdue social and traditional media. Especially social media remained a thorn in the government’s side. That’s why the government doggedly seeks to dilute freedom previously accrued to social media, the last platform that dissidents were supposed to enjoy unchallenged.
• Turkey’s democratic breakdown undercuts media freedoms, while social media appear the last lines of defense. Turkey’s pivot away from democratic path has only exacerbated the precarious state of media freedom. This leaves social media as the last contested battleground where critical voices speak out. The shifting international context would have a fallout over the next course of debate about the future of social media and political attempts to redesign it. Where Turkey would stand depends as much on its own political currents in Ankara as on the evolving state of world politics. (But this international aspect is out of the scope of this study, thus it is duly omitted.)

Conclusion

Our analysis of the 6,252 Turkish state-sponsored accounts suspended by Twitter shows the extent and modus operandi of the AKP government’s political astroturfing campaigns through social media during critical periods elaborated in a detailed fashion above. The main goals of these campaigns were to dominate public discourse, set the public agenda in accordance with the best interests of the AKP administration, and vilify and discredit political discontents.

The most remarkable method used by the troll accounts in these campaigns was masquerading as different political identities (or sockpuppet typologies) to disseminate their messages and false information among various segments of Turkish society and across the political spectrum. Based on self-identifications and the content of the tweets, we grouped the AKTroll accounts into four categories: pro-Erdogan, Kemalist, Nationalist, and pro-Kurdish. These four categories represent the major political groups in Turkey as they are currently represented in the Parliament. The troll accounts that did not fit into one of the four categories typically were ones that posed as the supporters of other political parties, the accounts of news agencies, and apolitical identities.

The combination of quantitative analysis (i.e., account creation dates and frequencies, variation among the activities of trolls in the different sockpuppet categories, and popular hashtags) and qualitative analysis (i.e., themes emerging from the tweets of top trolls) show how the AKP attempted to shape the public agenda through political astroturfing campaigns during elections and after critical national events.

The methods illustrated in the Discussion section became more visible during critical national events that unsettled the political contours in the past decade can be described as follows: the Gezi Park protests (2013), the corruption investigations (2013), the Kobani protests (2014), and the July 2016 coup attempt. In a concerted effort, the trolls’ Twitter campaigns portrayed the Gezi Park protests as a deliberate coup attempt to unseat the government via social upheaval across the nation, while it sought to project the sprawling December 2013 graft probes as a judicial coup in the same vein. Following the same line of logic, the trolls depicted the Kobani protests as the work of the Kurdish political opposition to instigate a violent social uprising against the government in the eastern parts of the country. A sweeping Twitter campaign was again in full swing after the abortive putsch in July 2016. The trolls presented what happened as an attempt by the Western proxies in Turkey to remove Erdogan’s government from power. Those groups, in collaboration with CIA, NATO, Pentagon, were behind the botched coup, so the trolls contended.

For the silenced opposition groups and the human rights defenders in Turkey, social media has been left as the last stand against the ever-expanding political encroachment over the boundaries of individual expression. It is the last viable platform where different people from all walks of life voice their grievances about the ongoing injustices and their personal plight in the country. Given the high Internet penetration rate in Turkey and the extensive use of social media by the tech-savvy Turkish people, the control of social media has become an imperative for the government, which deeply loathes any free room for dissent. Our findings confer credibility on the assessment that bears resemblance to a number of authoritarian governments such as Russia, China, and Iran in terms of political astroturfing campaigns in domestic realm.

As Turkey’s flagging economy forces an unpleasant reckoning for its rulers, the AKP’s popular support suffers an inexorable decline among its traditional constituency.^[59] This only adds to importance of the mediums of public debate. As social media re-asserts itself as the medium of choice for more and more Turkish people, it tempts the AKP rule to refine its subtle methods and to reconsider its campaigns on social media in an effort to define the public context of political debate there.

References

^[1] Stier, S., Bleier, A., Lietz, H., & Strohmaier, M. (2018, January). Election campaigning on social media: Politicians, audiences, and the mediation of political communication on Facebook and Twitter. Political Communication, 35(1), 50-74. doi:10.1080/10584609.2017.1334728

^[2] Keller et al. (2020), p. 1.

^[3] Ibid.

^[4] Twitter Safety. (2020, June 12). Disclosing networks of state-linked information operations we’ve removed. Safety Tools, Resources, and Updates.  https://blog.twitter.com/en_us/topics/company/2020/information-operations-june-2020.html

^[5] Hunter, I. (2015, June 6). Turkish president’s social media campaigners deny being a troll army. The Independent. http://www.independent.co.uk/news/world/europe/turkish-president-s-social-media-cam- paigners-deny-being-a-troll-army-10301599.html

^[6] Piers Brendon, “Death of Truth: When Propaganda and ‘Alternative Facts’ First Gripped the World,” the Guardian (The Guardian, March 11, 2017), https://www.theguardian.com/media/2017/mar/11/death-truth-propaganda-alternative-facts-gripped-world.

^[7] “‘Alternative Facts’ and the Costs of Trump-Branded Reality (Published 2017).” The New York Times, 2021. https://www.nytimes.com/2017/01/22/business/media/alternative-facts-trump-brand.html.

^[8]

European Commission. (2018). A multi-dimensional approach to disinformation: Report of the independent high-level group on fake news and online disinformation; p. 11. https://ec.europa.eu/digital-single-market/en/news/final-report-high-level-expert-group-fake-news-and-online-disinformation

^[9] Van Dijk, T. A. (2006). Discourse and manipulation. Discourse & Society, 17(2), 359-383; p. 359.

^[10] Ibid., p. 360.

^[11] Ibid.

^[12] Stier, S., Bleier, A., Lietz, H., & Strohmaier, M. (2018, January). Election campaigning on social media: Politicians, audiences, and the mediation of political communication on Facebook and Twitter. Political Communication, 35(1), 50-74. doi:10.1080/10584609.2017.1334728

^[13] Zimmermann, F., & Kohring, M. (2020). Mistrust, disinforming news, and vote choice: A panel survey on the origins and consequences of believing disinformation in the 2017 German parliamentary election. Political Communication, 37(2), 215-237. doi:10.1080/10584609.2019.1686095

^[14] Freelon, D., & Wells, C. (2020). Disinformation as political communication. Political Communication, 37(2), 145-156; p. 147. doi:10.1080/10584609.2020.1723755

^[15] Keller, F., Schoch, D., Stier, S., & Yang, J. (2019, October 28). It’s not easy to spot disinformation on Twitter: Here’s what we learned from 8 political “astroturfing” campaigns. The Washington Post. https://www.washingtonpost.com/politics/2019/10/28/its-not-easy-spot-disinformation-twitter-heres-what-we-learned-political-astroturfing-campaigns/; Lukito, J., Wells, C., Zhang, Y., Doroshenko, L., Kim, S. J., Su, M.-H., Suk, J., Xia, Y., & Freelon, D. (2018). The Twitter exploit: How Russian propaganda infiltrated U.S. news. Madison, WI: University of Wisconsin-Madison Social Media and Democracy Research Group. https://uwmadison.app.box.com/v/TwitterExploit

^[16] Keller et al. (2020), p. 1.

^[17] Ibid.

^[18] Ibid.

^[19] Guo, L., Mays, K., & Wang, J. (2019). Whose story wins on Twitter? Journalism Studies, 20(4), 563-584. doi:10.1080/1461670X.2017.1399813

^[20] McCombs, M. E., & Shaw, D. L. (1972). The agenda-setting function of mass media. Public Opinion Quarterly, 36(2), 176-187. doi:10.1086/267990

^[21] Guo, L., & McCombs, M. (2015). The power of information networks: New directions for agenda setting. New York, NY: Routledge.

^[22] Keller et al. (2019).

^[23] Freelon, D., Bossetta, M., Wells, C., Lukito, J., Xia, Y., Adams, K., (2020). Black Trolls Matter: Racial and Ideological Asymmetries in Social Media Disinformation. Soc. Sci. Comput. Rev. https://doi.org/10.1177/0894439320914853

^[24] Twitter Safety. (2020, June 12). Disclosing networks of state-linked information operations we’ve removed. Safety Tools, Resources, and Updates.  https://blog.twitter.com/en_us/topics/company/2020/information-operations-june-2020.html

^[25] Grossman, S., Akis, F. A., Alemdaroğlu, A., Goldstein, J. A., Jonsson, K., Garcia-Camargo, I., Thiel, D., & Zaheer, A. (2020). Political retweet rings and compromised accounts: A Twitter influence operation linked to the youth wing of Turkey’s ruling party [White paper]. Stanford University Freeman Spogli Institute for International Studies.  https://fsi.stanford.edu/publication/june-2020-turkey-takedown

^[26] Saka, E. (2018). Social media in Turkey as a space for political battles: AKTrolls and other politically motivated trolling. Middle East Critique, 27(2), 161-177. doi:10.1080/19436149.2018.1439271

^[27] Bilgin, H. D. (2008). Foreign policy orientation of Turkey’s pro-Islamist parties: A comparative study of the AKP and Refah. Turkish Studies 9(3), 407-421.

^[28] Friedman, D. (2015, August 25). The Turkish model: The history of a misleading idea. Washington, DC: Center for American Progress. https://www.americanprogress.org/issues/security/reports/2015/08/25/119932/the-turkish-model/

^[29] Shibley Telhami. “The 2011 Arab Public Opinion Poll.” Brookings. Brookings, November 21, 2011. https://www.brookings.edu/research/the-2011-arab-public-opinion-poll/

^[30] Edelman, E., Cornell, S., Lobel, A., & Karaveli, H. (2015, October). Turkey transformed: The origins and evolution of authoritarianism and Islamization under the AKP. Washington, DC: Bipartisan Policy Center. https://bipartisanpolicy.org/wp-content/uploads/2019/03/BPC-Turkey-Transformed.pdf

^[31] Levitsky, S., & Way, L. A. (2002). Elections without democracy: The rise of competitive authoritarianism. Journal of Democracy, 13(2), 51-65.

^[32] San, S., & Akca, D. (2020). How Turkey’s democratic backsliding compromises international dimension of democratization? Digest of Middle East Studies, 30(1), 34-52.

^[33] Ozeren, S., & Cubukcu, S. (April 22, 2020). Coronavirus: The disastrous consequences of the Turkish media’s transformation. Jerusalem Post. https://www.jpost.com/opinion/coronavirus-the-disastrous-consequences-of-turkish-medias-transformation-625518

^[34] Keller, F. B., Schoch, D., Stier, S., & Yang, J. (2020). Political astroturfing on Twitter: How to coordinate a disinformation campaign. Political Communication, 37(2), 256-280; p. 1. doi:10.1080/10584609.2019.1661888

^[35] Ozeren, S., & Cubukcu, S. (2020, April 22). Coronavirus: The disastrous consequences of Turkish media’s transformation. The Jerusalem Post. https://www.jpost.com/opinion/coronavirus-the-disastrous-consequences-of-turkish-medias-transformation-625518

^[36] Hunter, I. (2015, June 6). Turkish president’s social media campaigners deny being a troll army. The Independent. http://www.independent.co.uk/news/world/europe/turkish-president-s-social-media-cam- paigners-deny-being-a-troll-army-10301599.html

^[37] Saka. (2018).

^[38] Andı, S., Aytaç, E., & Carkoglu, A. (2019, July). Internet and social media use and political knowledge: Evidence from Turkey. Mediterranean Politics, 25(5), 579-599. doi:10.1080/13629395.2019.1635816

^[39] O’Donohue, A., Hoffman, M., & Makovsky, A. (2020, June 10). Turkey’s changing media landscape. Washington, DC: Center for American Progress.  https://www.americanprogress.org/issues/security/reports/2020/06/10/485976/turkeys-changing-media-landscape/

^[40] Heinrich, D. (2020, October 4). User safety or censorship? Turkey targets social media platforms. Deutsche Welle. https://www.dw.com/en/user-safety-or-censorship-turkey-targets-social-media-platforms/a-55150477

^[41] Saka. (2018).

^[42] Saka. (2018).

^[43] Amnesty International. (2019, April 30). Turkey: First academic to go to prison for signing peace petition in a flagrant breach of freedom of expression. https://www.amnesty.org/download/Documents/EUR4402902019ENGLISH.pdf

^[44] Human Rights Watch. (2018, October 17). Turkey: End prosecutions for “insulting president”: Students on trial despite right to free speech. https://www.hrw.org/news/2018/10/17/turkey-end-prosecutions-insulting-president

^[45] Freedom House. (2020, July 30). Turkey: Passage of social media law curtails human rights online [Press release]. https://freedomhouse.org/article/turkey-passage-social-media-law-curtails-human-rights-online

^[46] Ibid.

^[47] Information operations: About this report. (n.d.). Twitter.  https://transparency.twitter.com/en/reports/information-operations.html. We have fewer accounts (6,252) than the Stanford report does (7,340).

^[48] Ibid.

^[49] Ibid.

^[50] Albayrak, A. and Parkinson, J. (2013, September 16). Turkey’s Government Forms 6,000-Member Social Media Team. Wall Street Journal. https://online.wsj.com/article/SB10001424127887323527004579079151479634742.html

^[51] Benedictus, L. (2016, November 6). Invasion of the troll armies: from Russian Trump supporters to Turkish state stooges. The Guardian. https://www.theguardian.com/media/2016/nov/06/troll-armies-social-media-trump-russian

^[52] Daily Sabah (2015, May 11). AK Party founded New Turkey Digital Office for the general elections on June 7. https://www.dailysabah.com/elections/2015/05/11/ak-party-founded-new-turkey-digital-office-for-the-general-elections-on-june-7

^[53] Sozcu (2015, September 29).  Aktroller ne kadar maaş alıyor?  https://www.sozcu.com.tr/2015/gunun-icinden/aktroller-ne-kadar-maas-aliyor-945937/

^[54] Hurriyet Daily News (2015, June 15). Erdoğan attends ‘‘Ak troll’ wedding, chats with well-known suspect. https://www.hurriyetdailynews.com/erdogan-attends-ak-troll-wedding-chats-with-well-known-suspect-84013

^[55] Cumhuriyet (2020, June 16). CHP’nin Millet Dergisi’nden ‘Troll Özel’ sayısı: Kim bu aktroller?  https://www.birgun.net/haber/chp-nin-millet-dergisi-nden-troll-ozel-sayisi-kim-bu-aktroller-304824

^[56] Dailymotion (2019). Mansur Yavaş: Hayvanlara Dindar İnsanların Çocukları Daha Fazla Eziyet Ediyor https://www.dailymotion.com/video/x7175pu

^[57] BBC (2015, May 13). Turkey mine disaster: Raw anger in Soma a year on. https://www.bbc.com/news/world-europe-32709431

^[58] Perry, B., Akca, D., Karakus, F., & Bastug, M. F. (2020). Planting hate to harvest hatred: How does the political hate speech fuel hate crimes against targeted groups? International Journal for Crime, Justice and Social Democracy, 9(4), 195-211. doi:10.5204/ijcjsd.v9i4.1514

^[59] MetroPOLL. [@metropoll]. (2020, November 9). Türkiye’nin Nabzı Ekim 2020: Bu Pazar milletvekili seçimi olsa hangi siyasi partiye oy verirsiniz? [Pulse of Turkey in October 2020: Which political party do you vote for if parliamentary elections this Sunday?] [Tweet]. Twitter.    https://twitter.com/metropoll/status/1325712718662410241

______________________________________________________

Orion Policy Institute (OPI) is an independent, non-profit, tax-exempt think tank focusing on a broad range of issues at the local, national, and global levels. OPI does not take institutional policy positions. Accordingly, all views, positions, and conclusions represented herein should be understood to be solely those of the author(s) and do not necessarily reflect the views of OPI.