Explore through two lenses
SR/EN
AUTHORITARIANISM - RESEARCH ATLASAVtokratija

How the authoritarian system works in Serbia

General research connected to documented examples, findings, and sources from Serbia.

100questions
20areas
148academic works
130Serbia sources
Go to the general AUTLAS Explore general mechanisms of authoritarian systems without the Serbia evidence layer.
→

SIX ENTRIES INTO THE MATRIX

Which question do you want to start with?

Concrete questions are the starting point. The other 94 are available through the map of 20 areas.

MAP OF THE AUTHORITARIAN SYSTEM

Twenty areas guide you through the system as a whole.

Select an area to see why it matters and which five questions it opens.

AVTOKRATIJA›How digital surveillance and bad news change power
57
Research-based answerAUTLAS Research Base

How do bots, organized networks, and algorithms make a message seem larger and more popular than it is?

Confidence in the conclusionquite substantiated, with important conditions
key academic works5
linked Serbia sources4
Period201226

Bots and organized networks can give the same message the appearance of mass and spontaneous support. A large number of orders repeat the same claims, artificially raise the mark among popular topics, attack critics and bury the conversation so that the actual participants withdraw. The platform algorithm can then further spread content that already receives a lot of reactions, regardless of whether they are authentic.

The important effect is not just persuasion, but creating the impression that “everybody thinks so”, that opposition is lonely or that public speaking will bring an attack. However, high visibility is not the same as changing opinions, and bot detection tools are wrong. The claim seeks evidence of a harmonized time, content and account network, as well as the careful separation of an organized campaign from actual popularity.

Example from SerbiaChecked connection 4 sources; waiting for saturation pass
4 related sources; Serbia not yet logged on Saturation Pass

Organized accounts, party activists and algorithms can make the same message seem like the spontaneous mood of thousands of people. In Serbia, network accounts and coordinated commenting have been repeatedly linked to the promotion of power and attacks on critics, while television dominance spills over into social networks. The number of repetitions does not only measure support; it measures organizational money, access and the ability to take up space.

How the finding is separated in the matrix

01 Established fact

Domestic and international research records coordinated pro-government networks and digital harassment, and media monitoring shows content that such networks can further amplify.

02 Our interpretation

The algorithm rewards attention, so organized initial activity can attract real users and turn an artificial signal into a real political effect.

03 Does it show the same process?

To a significant extent coordination has been documented on several occasions, but the identity and ordering party of each order is not known.

04 Important caveat or alternative explanation

The popular message is not automatically the result of bots, and the party activist is not necessarily a fake account. Patterns of time, content, network connections and behavior are required to assert coordination.

What the general answer does not allow us to conclude

Programs that recognize bots and organized networks can be wrong. The high visibility of the message also does not prove that people have changed their minds.

Works supporting this answer, with their evidence type and an explanation of how they relate to the claim.

A01

Explain how the system works

Schlumberger, O., Edel, M., Maati, A. i Saglam, K. (2024), „How Authoritarianism Transforms: A Framework for the Study of Digital Dictatorship“, Government and Opposition 59(3): 761–783.

Contribution to this answer
The paper shows how digital technologies can increase the ability of authoritarian authorities to monitor, direct and shape political communication, but it does not specifically study how bots or coordinated networks artificially increase the popularity of a message. Here is a wider framework of digital authoritarianism.
Open source ↗
A02

research that checks more closely what the cause is

Milli, S. i dr. (2025), „Engagement, user satisfaction, and the amplification of divisive content on social media“, PNAS Nexus 4(3): pgaf062.

Contribution to this answer
Milli and co-authors show very concrete what the algorithm itself can do: ranking according to engagement gives more visibility to emotionally charged and politically divisive content than a chronological representation, even when users themselves do not say that they want such content more. The message may therefore seem more important simply because the system rewards it with a greater range.
Open source ↗
A03

research that checks more closely what the cause is

Brady, W. J., McLoughlin, K., Doan, T. N. i Crockett, M. J. (2021), „How Social Learning Amplifies Moral Outrage Expression in Online Social Networks“, Science Advances 7(33): eabe5641.

Brady, W. J., McLoughlin, K., Doan, T. N. i Crockett, M. J. (2021), "How Social Learning Amplifies Moral Outrage Expression in Online Social Networks", Science Advances 7(33): eabe5641.

Contribution to this answer
Brady and the co-authors show another mechanism of reinforcement: when the expression of moral indignation receives more social rewards likes, shares and other reactions people learn to repeat it more often. Thus, the network itself can produce more and more content that looks like a broad and intense mood of the public.
Open source ↗
A04

study of Iran and international technology firms

Conduit, D. (2025), „Digital Authoritarianism and the Global Technology Industry: Evidence from Iran“, Government and Opposition 60(3): 751–772.

Contribution to this answer
Conduit shows how global technology companies can enable authoritarian regimes to have sophisticated digital control and operations, but the work is not about bots, algorithmic rankings or artificial creation of popularity messages.
Open source ↗
A05

review more research

Roberts, M. E. (2020), „Resilience to Online Censorship“, Annual Review of Political Science 23: 401–419.

Contribution to this answer
Roberts studies when people manage to bypass Internet censorship and give up on it. This is important for digital information space, but does not explain the mechanism of bots, coordinated amplification or algorithmic popularity.
Open source ↗

Sources supporting the Serbia example, with the claim they support and their limitations.

S01

Technology Platform Report Technical Attribution of the Network - Verified 21.08.2026

Nimmo, Ben, Margarita Franklin, David Agranovich, Lindsay Hundley i Mike Torrey. 2023. Meta Quarterly Adversarial Threat Report, Q4 2022. Menlo Park: Meta.

What the source records
The target has removed 5,374 Facebook accounts, 12 groups and 1,060 Instagram accounts from a network that has created the appearance of widespread spontaneous support for the president and the SNS and criticized the opposition.
Claim it supports
The claim that coordinated orders and the media ecosystem can create the illusion of mass spontaneous support and intensify the attack on critics.
Source limitation
The platform sees only the activities on its own services and publishes the conclusions of the internal investigation, without a complete set of data that independent researchers could repeat as a whole; the findings do not measure the conclusiveness or electoral performance of the network.
Open source ↗
S02

Domestic NGOs systematic media monitoring - Verified 21.08.2026

CRTA. 2025. Media Monitoring of Political Pluralism: Government in Focus, Opposition Almost Invisible. Beograd: CRTA.

What the source records
Monitoring finds that more than half of the time intended for political actors went to the president, who had 354 live television broadcasts and almost exclusively neutral or positive treatment.
Claim it supports
The claim that coordinated orders and the media ecosystem can create the illusion of mass spontaneous support and intensify the attack on critics.
Source limitation
The sample includes the central news of five national TV stations, but does not measure all media, the entire program offer, the audience or the direct impact of such reporting on voting.
Open source ↗
S03

Reviewed scientific work - Verified 21.08.2026

Dragojlov, Aleksandra. 2026. “Influence of Political Clientelism on Media Freedom under Vučić and the Progressive Party.” Nationalities Papers 54(2): 391–407. https://doi.org/10.1017/nps.2025.25

What the source records
The paper concludes that subtle mechanisms of indirect control undermined media freedom despite the formal adoption of new laws.
Claim it supports
The claim that coordinated orders and the media ecosystem can create the illusion of mass spontaneous support and intensify the attack on critics.
Source limitation
It is focused on the media sector and relies partly on a limited number of professional interviews.
Open source ↗
S04

Domestic NGO monitoring of civic space - Verified 21.08.2026

Građanske inicijative. 2025. Udruženja građana — sužavanje prostora za delovanje 2024. Beograd: Građanske inicijative.

What the source records
Based on the standardized regional matrix and bi-weekly reports, 234 cases of violations of fundamental rights and freedoms were registered, with an increase in physical and verbal attacks and other pressures compared to the previous year.
Claim it supports
The claim that coordinated orders and the media ecosystem can create the illusion of mass spontaneous support and intensify the attack on critics.
Source limitation
The records include documented and reported cases, and not a representative rate of all incidents in Serbia, while the organization at the same time has a clear role of protecting the civil space.
Open source ↗

ABOUT US AND OUR METHODOLOGY

Many sources. Clear methodology. Verifiable answers.

AUTLAS is a public knowledge platform about how authoritarian systems work. It turns research into clear answers about elections, media, institutions, public resources, pressure, and everyday life.

01

Based on academic research

Answers draw on academic research on authoritarian systems, democratic backsliding, institutional capture, and the narrowing of civic rights and freedoms.

02

Practice documented in Serbia

AVtokratija is the Serbia case study within AUTLAS. It uses reports, analyses, and data from relevant organizations, institutions, and investigative media. General findings and Serbia evidence are presented separately.

03

Sources open to scrutiny

Each answer lists its sources so visitors can check the basis for its conclusions.

04

Why questions?

AUTLAS organizes dispersed knowledge through 100 questions in 20 areas, helping people understand the wider system behind familiar phenomena.

05

Focus on the system

AUTLAS analyzes rules, incentives, pressures, and institutional mechanisms.

06

Who is behind AUTLAS?

AUTLAS and AVtokratija were conceived and are developed by Civic Initiatives. Content is updated in line with new research and reliable data.

CONTINUOUSLY UPDATED

Help us keep the content complete and accurate.

Send a relevant source, point out an error, or suggest how an explanation could be clearer. Every submission is reviewed before the content is changed.

Send a source, report an error, or share a suggestion