Distrust in the state is widespread among young South Africans, contributing to polarisation and raising urgent questions for democratic renewal. Distrust in the state is widespread among young South Africans and is contributing to a serious democratic deficit. Amid a new wave of anti-immigrant protests, and with local elections coming up in November 2026, investments in dialogue are urgently needed to address levels of polarisation.
Youth, social media and affective polarisation
The project “Youth, Affective Polarisation and Trust” (YAPT) focuses on the role of social media engagement in dynamics of affective polarisation in South Africa, Brazil and India.
Affective polarisation, when supporters of opposing political parties are not just disagreeing on issues but also feeling strong antipathy towards one another, is affecting democracies around the globe. It is one feature of democratic backsliding, and may drive backsliding if citizens elect incapable, corrupt or even autocratic political leaders simply because they see them as part of “their” camp.
Social media plays a strong role in the circulation of polarising narratives. A parallel trend that has unfolded is that young people have less faith in democracy than adult generations. How do these issues connect? This question is addressed in the YAPT project, which started in September 2024.
The contexts of the three YAPT focus countries are quite different. Polarisation in Brazil and India is largely driven by two main “camps”: those for and against the Workers’ Party (Partido dos Trabalhadores) in <a href="https://www.cambridge.org/core/journals/latin-american-politics-and-society/article/partisan-stereotyping-and-polarization-in-brazil/2F11FF3E4FB1C4B37069FA05515F29DA” rel=”nofollow noopener” target=”_blank”>Brazil, and the supporters and adversaries of the Hindu-nationalist Bharatiya Janata Party (BJP) in India.
In South Africa, on the other hand, polarisation is not binary due to the role of multiple parties. The country is characterised by high levels of social polarisation, rooted in its apartheid legacy and shaped by intersecting racial, socioeconomic and political inequalities. While the project was underway, dynamics in South Africa unfolded whereby pre-existing negative sentiments targeting undocumented immigrants deepened.
Political parties like the Patriotic Alliance, ActionSA and uMkhonto we Sizwe have promoted narratives that blame undocumented immigrants from other African countries for taking jobs and using public services, capitalising on grievances over inequalities, lack of decent housing and employment prospects.
The March and March campaign has organised nationwide anti-immigrant protests and demanded that undocumented immigrants leave by 30 June. The movement has sparked actions that President Cyril Ramaphosa called “vigilantism”, with mobs attacking immigrant homes and shops and telling African nationals to leave, including those who reside in the country legally.
Meanwhile, young people largely feel disenfranchised from politics, and many abstained from voting in the 2024 elections. Many express their discontent online, where they encounter heated and polarising debates. How, then, do young South Africans engage in these debates?
Distrust in the South African state
Our initial analysis, which combines qualitative data with social media analysis, shows that broader patterns of polarisation and identity-based tensions are connected to a deep lack of trust in state institutions, a trend identified in previous studies about growing distrust in political party leadership.
While concerns about political leadership are not new, the volume of online discussions suggests that these concerns are becoming increasingly central to public debate.
YAPT analysed social media posts between January and September 2025 using Large Language Models (LLMs). Distrust in political party leadership emerged as one of the most prominent themes, generating thousands of comments and making it one of the most actively discussed political issues.
The findings reveal a rise in political party polarisation and antagonism, as well as racial, ethnic and national identity tensions around February 2025. This period coincided with South Africa’s hosting of the G20 Foreign Ministers’ Meeting in Johannesburg. The event took place amid diplomatic tensions between South Africa and the United States over US policy changes, South Africa’s position on the Middle East, and its land policies. Such developments fostered further polarising narratives around identity and political allegiance.
Online comments reflecting distrust in government spiked right after President Cyril Ramaphosa’s suspension of Minister of Police Senzo Mchunu, following allegations linking the minister to organised crime networks. Although the matter is still under investigation, such allegations strike at the heart of public confidence.
The debate spiked again following the announcement of the National Dialogue process in June, a government initiative intended to promote social cohesion, trust, nation-building and democratic renewal. However, distrust in political leadership raises questions about accompanying efforts to strengthen social cohesion.
In focus groups, young people also talked about corruption of the state as a main cause of the country’s challenging conditions. A fundamental cause of weak service delivery in Soweto, a large township by Johannesburg where we spoke with youth, is the spatial segregation and marginalisation caused by apartheid.
Discontent has driven protest and violent action over the years, and is now targeting undocumented migrants rather than root causes. The social media analysis revealed far more posts showing distrust based on political parties than on anti-foreigner sentiments.
Studying online and offline dynamics
The YAPT project involves close collaboration between social scientists who speak to youth in all three countries and data scientists who use LLMs to conduct the social media analysis, working with data made available through Bright Data’s data-for-good initiative. The study made clear why this interdisciplinary work is essential.
Understanding youth affective polarisation requires more than analysing what is said online: it requires understanding who is saying it, and why. A key limitation of LLM-based analysis is that it cannot reliably verify the age of those posting or commenting. Platforms do not consistently surface this information, and young people may not self-identify as such online.
Dynamics observed on social media therefore had to be checked against what young people say offline and their lived experience. This informed the iterative research design. Researchers in South Africa asked youth to identify themes on social media they considered polarising. At the same time, we needed a theoretical understanding of how platforms shape polarisation among young users.
This combination of country-specific grounding and platform theory informed our analytical approach to understanding mechanisms and consequences: from what content says, to how polarisation occurs through specific platform features and processes, to the emotional intensity with which it registers as affective polarisation, to its effects on trust and youth engagement, and finally to what behavioural consequences follow.
An initial set of social media posts was labelled, and labels were reviewed, refined and validated by social scientists before being applied to classify the full corpus. Subsequent layers of LLM-assisted classification identified the forms, intensity and targets of emotion expressed in posts, and then detected expressions of distrust.
These classifications were integrated into a dashboard, allowing the team to explore patterns across time, topic, emotional register and target. This shows which topics are considered more polarising.
Moving forward
Collectively, our findings suggest that identity-based tensions do not exist in isolation. They appear closely connected to broader patterns of distrust in political leadership.
The challenge facing South Africa, therefore, is not only addressing political distrust but also strengthening democratic trust in a year of hotly contested local government elections due in November 2026.
In the next stage of the YAPT project, the team will examine whether structured dialogue, in the form of MiniPublics, can mitigate polarisation.
Acknowledgements
We thank the Bright Initiative by Bright Data for providing access to public social media data for the purpose of this project.
Disclaimer
The views expressed in this opinion piece are those of the author/s and do not necessarily reflect the views or policies of IDS.
