08 September 2025
- RSIS
- Publication
- RSIS Publications
- Battling Extremism: What Counts as Knowledge
SYNOPSIS
Recent years have shown that extremist worldviews are no longer limited to specific ideologies, regions, or grievances. Whether driven by Islamist militancy, far-right conspiracies, or historical grievances, the dissemination of radical beliefs today is shaped by a deeper and more fundamental issue: extremist epistemology, especially regarding how individuals come to know, filter, and reinforce what they believe to be true.

COMMENTARY
At the heart of radicalisation lies an epistemology that rigidly filters information, dismisses contradictory evidence, and resists alternative perspectives. Quassim Cassam and Olivia Bailey have described this as a “closed-minded worldview” that replaces open inquiry with ideological purity. In this view, epistemology refers not to formal theories of knowledge, but to the everyday frameworks and practices by which individuals justify their beliefs and decide what to believe or reject.
When a person becomes epistemically closed off – often through social media echo chambers or ideological networks – their epistemic autonomy is compromised. They no longer analyse evidence critically, instead relying on trusted sources or in-groups to determine what qualifies as “truth.” This vulnerability is what extremist groups exploit, online and offline.
Self-radicalised individuals – including those in Singapore detained under the Internal Security Act for plotting attacks or attempting to travel to conflict zones – often fell into these epistemic traps. Although the content they consumed may have varied, their radicalisation process was similar: they entered epistemic environments that made violence appear not only justified but also necessary.
These environments often revolve around radical ideologies that reinforce extremist epistemology by offering binary moral frameworks that simplify complex realities and by providing emotionally resonant certainties about the future – such as promises of martyrdom or apocalyptic triumph. These approaches help define individuals as they seek clarity, purpose, or control amid uncertainty.
A Global Pattern of Epistemic Closure
This pattern is not unique to any particular country or ideology. Across various contexts, extremist epistemology has manifested in troubling and violent ways.
In the United Kingdom, Jaswant Singh Chail was sentenced to nine years in prison after attempting to assassinate Queen Elizabeth II in 2021. Chail cited revenge for the 1919 Jallianwalah Bagh massacre as his motive. In court, it was revealed that he had spent months conversing with an AI chatbot named “Sarai” on the Replika app. The chatbot reportedly affirmed his violent intentions, feeding into a loop of self-radicalisation that blurred reality and fantasy.
In Germany, the dismantling of a radical Reichsbürger network in 2022 revealed how conspiratorial worldviews, obsessed with restoring a pre-democratic state, could fuel real-world terror plots. Although the plot was disrupted in its early stages, it demonstrated how belief in disinformation and epistemic isolation could lead to organised attempts at overthrowing a government.
In Indonesia, despite the absence of large-scale attacks in 2023-24, authorities warn that radicalisation pipelines remain active – especially among youths. The National Counterterrorism Agency (BNPT) and the Ministry of Communication and Information removed over 180,000 pieces of radicalism-related content in 2024 alone. Analysts caution that terrorist networks are shifting from coordinated attacks to decentralised online recruitment and mobilisation.
The common denominator across these cases is epistemic. Extreme ideologies tend to thrive under specific conditions, such as when critical questioning is discouraged, opposing viewpoints are ignored or rejected, and individuals engage solely with like-minded peers. These dynamics foster an echo chamber that shields extreme beliefs from scrutiny, allowing them to flourish unchecked and gain traction without the moderating influence of diverse perspectives.
Digital Platforms and AI: Multipliers of Extremist Knowing
Social media platforms play a significant role in shaping epistemic environments. Singapore’s cases of self-radicalised individuals highlight how online spaces serve as accelerators for extremist thought and ideas. These platforms often present religious and political content in emotionally charged formats, stripped of context, and optimised for engagement, not accuracy.
Three cases involving Singaporean youths, aged 14 to 21, highlight the worrying issue of online self-radicalisation. Influenced by ISIS propaganda and far-right content on platforms like YouTube and extremist forums, they progressed from consuming hateful material to actively planning attacks against mosques and ethnic communities in Singapore, demonstrating how digital echo chambers can quickly turn extremist ideologies into violent extremism.
Artificial Intelligence compounds the problem. Chatbots, algorithmic curation, and deepfake content can create personalised, scalable systems of ideological affirmation. An analysis by the International Centre for Counter-Terrorism (ICCT) based in the Netherlands has warned of AI’s dual potential: on the one hand, enabling recruitment, micro-targeting, and disinformation; on the other, providing tools for counter-radicalisation and simulation of interventions.
In Denmark, intelligence authorities have noted how AI-generated content circulates through memes and coded language, making it more difficult to detect and increasing the likelihood of reaching vulnerable young people. Tech Against Terrorism, the UN-backed international initiative to fight terrorist activity within online technology, archived over 5,000 pieces of AI-generated extremist content in 2023, revealing the scale of experimentation underway.
However, it is important to note that exposure alone does not necessarily turn someone into an extremist. Radicalisation occurs when individuals become epistemically locked in, relying solely on one narrative and rejecting counterevidence. When people surrender their epistemic autonomy to an ideology or group, they no longer decide what is true through reasoning – truth becomes whatever the group affirms.
This explains why, despite credible religious education and interfaith outreach in Singapore and elsewhere, some individuals still fall into radicalisation pathways. It is not the absence of credible information that leads to radicalisation, but the dominance of epistemic spaces that suppress doubt and close off engagement with a plurality of views.
Policy Imperatives: Rebuilding Epistemic Resilience
Countering radicalisation requires more than simply deleting content or deplatforming users. It demands a restructuring of the epistemic environments where beliefs are formed and contested.
First, education systems must incorporate digital and critical thinking skills into school curricula. Students should learn how to evaluate sources, detect AI-generated manipulation, and resist confirmation bias.
Second, community resilience matters. Religious, civic, and cultural leaders should be equipped to foster spaces where empathy, dialogue, and open inquiry are encouraged – not just in response to extremism, but as a regular part of life.
Third, platform accountability is critical. Social media and technology companies must extend their moderation efforts to encrypted and fringe spaces, with particular vigilance toward AI-generated content that blurs the boundaries between fiction and reality.
Fourth, international cooperation must keep pace with technological and ideological change. Cross-border efforts, such as Indonesia’s cybersecurity partnership with the Netherlands, provide models for transnational epistemic resilience.
Finally, rehabilitation and reintegration efforts remain vital. Programmes in Indonesia, Singapore, and across Europe show that epistemic change is possible – when personal reflection, community support, and state intervention come together.
Conclusion: The Battle Over How We Know
At its core, the challenge of radicalisation revolves around how people acquire knowledge. Whether in London, Jakarta, Berlin, or Singapore, the same pattern unfolds: grievance is curated online, epistemic closure hardens, and violence becomes plausible.
To confront this, we must look beyond mere content removal or censorship and address the epistemic systems that undergird extremist beliefs. Only by fostering openness, critical inquiry, and mutual understanding – both online and offline – can we reshape the environments in which dangerous ideologies take root.
In an age of AI and algorithmic persuasion, the battle against extremist epistemology is not just about where information is obtained or what people believe, but about how they learn to believe it.
About the Authors
Mohamed Bin Ali is a Senior Fellow in the Inter-Religious Relations in Plural Societies Programme, and Muhammad Haziq Bin Jani is a Senior Analyst in the Indonesia Programme at the S. Rajaratnam School of International Studies (RSIS), Nanyang Technological University (NTU). Sabariah Hussin holds a PhD from the Department of Religion at Rice University in the USA. She is also an alumnus of RSIS. Mohamed Bin Ali is the co-Chairman and Sabariah Hussin a member of the Religious Rehabilitation Group (RRG), Singapore.
SYNOPSIS
Recent years have shown that extremist worldviews are no longer limited to specific ideologies, regions, or grievances. Whether driven by Islamist militancy, far-right conspiracies, or historical grievances, the dissemination of radical beliefs today is shaped by a deeper and more fundamental issue: extremist epistemology, especially regarding how individuals come to know, filter, and reinforce what they believe to be true.

COMMENTARY
At the heart of radicalisation lies an epistemology that rigidly filters information, dismisses contradictory evidence, and resists alternative perspectives. Quassim Cassam and Olivia Bailey have described this as a “closed-minded worldview” that replaces open inquiry with ideological purity. In this view, epistemology refers not to formal theories of knowledge, but to the everyday frameworks and practices by which individuals justify their beliefs and decide what to believe or reject.
When a person becomes epistemically closed off – often through social media echo chambers or ideological networks – their epistemic autonomy is compromised. They no longer analyse evidence critically, instead relying on trusted sources or in-groups to determine what qualifies as “truth.” This vulnerability is what extremist groups exploit, online and offline.
Self-radicalised individuals – including those in Singapore detained under the Internal Security Act for plotting attacks or attempting to travel to conflict zones – often fell into these epistemic traps. Although the content they consumed may have varied, their radicalisation process was similar: they entered epistemic environments that made violence appear not only justified but also necessary.
These environments often revolve around radical ideologies that reinforce extremist epistemology by offering binary moral frameworks that simplify complex realities and by providing emotionally resonant certainties about the future – such as promises of martyrdom or apocalyptic triumph. These approaches help define individuals as they seek clarity, purpose, or control amid uncertainty.
A Global Pattern of Epistemic Closure
This pattern is not unique to any particular country or ideology. Across various contexts, extremist epistemology has manifested in troubling and violent ways.
In the United Kingdom, Jaswant Singh Chail was sentenced to nine years in prison after attempting to assassinate Queen Elizabeth II in 2021. Chail cited revenge for the 1919 Jallianwalah Bagh massacre as his motive. In court, it was revealed that he had spent months conversing with an AI chatbot named “Sarai” on the Replika app. The chatbot reportedly affirmed his violent intentions, feeding into a loop of self-radicalisation that blurred reality and fantasy.
In Germany, the dismantling of a radical Reichsbürger network in 2022 revealed how conspiratorial worldviews, obsessed with restoring a pre-democratic state, could fuel real-world terror plots. Although the plot was disrupted in its early stages, it demonstrated how belief in disinformation and epistemic isolation could lead to organised attempts at overthrowing a government.
In Indonesia, despite the absence of large-scale attacks in 2023-24, authorities warn that radicalisation pipelines remain active – especially among youths. The National Counterterrorism Agency (BNPT) and the Ministry of Communication and Information removed over 180,000 pieces of radicalism-related content in 2024 alone. Analysts caution that terrorist networks are shifting from coordinated attacks to decentralised online recruitment and mobilisation.
The common denominator across these cases is epistemic. Extreme ideologies tend to thrive under specific conditions, such as when critical questioning is discouraged, opposing viewpoints are ignored or rejected, and individuals engage solely with like-minded peers. These dynamics foster an echo chamber that shields extreme beliefs from scrutiny, allowing them to flourish unchecked and gain traction without the moderating influence of diverse perspectives.
Digital Platforms and AI: Multipliers of Extremist Knowing
Social media platforms play a significant role in shaping epistemic environments. Singapore’s cases of self-radicalised individuals highlight how online spaces serve as accelerators for extremist thought and ideas. These platforms often present religious and political content in emotionally charged formats, stripped of context, and optimised for engagement, not accuracy.
Three cases involving Singaporean youths, aged 14 to 21, highlight the worrying issue of online self-radicalisation. Influenced by ISIS propaganda and far-right content on platforms like YouTube and extremist forums, they progressed from consuming hateful material to actively planning attacks against mosques and ethnic communities in Singapore, demonstrating how digital echo chambers can quickly turn extremist ideologies into violent extremism.
Artificial Intelligence compounds the problem. Chatbots, algorithmic curation, and deepfake content can create personalised, scalable systems of ideological affirmation. An analysis by the International Centre for Counter-Terrorism (ICCT) based in the Netherlands has warned of AI’s dual potential: on the one hand, enabling recruitment, micro-targeting, and disinformation; on the other, providing tools for counter-radicalisation and simulation of interventions.
In Denmark, intelligence authorities have noted how AI-generated content circulates through memes and coded language, making it more difficult to detect and increasing the likelihood of reaching vulnerable young people. Tech Against Terrorism, the UN-backed international initiative to fight terrorist activity within online technology, archived over 5,000 pieces of AI-generated extremist content in 2023, revealing the scale of experimentation underway.
However, it is important to note that exposure alone does not necessarily turn someone into an extremist. Radicalisation occurs when individuals become epistemically locked in, relying solely on one narrative and rejecting counterevidence. When people surrender their epistemic autonomy to an ideology or group, they no longer decide what is true through reasoning – truth becomes whatever the group affirms.
This explains why, despite credible religious education and interfaith outreach in Singapore and elsewhere, some individuals still fall into radicalisation pathways. It is not the absence of credible information that leads to radicalisation, but the dominance of epistemic spaces that suppress doubt and close off engagement with a plurality of views.
Policy Imperatives: Rebuilding Epistemic Resilience
Countering radicalisation requires more than simply deleting content or deplatforming users. It demands a restructuring of the epistemic environments where beliefs are formed and contested.
First, education systems must incorporate digital and critical thinking skills into school curricula. Students should learn how to evaluate sources, detect AI-generated manipulation, and resist confirmation bias.
Second, community resilience matters. Religious, civic, and cultural leaders should be equipped to foster spaces where empathy, dialogue, and open inquiry are encouraged – not just in response to extremism, but as a regular part of life.
Third, platform accountability is critical. Social media and technology companies must extend their moderation efforts to encrypted and fringe spaces, with particular vigilance toward AI-generated content that blurs the boundaries between fiction and reality.
Fourth, international cooperation must keep pace with technological and ideological change. Cross-border efforts, such as Indonesia’s cybersecurity partnership with the Netherlands, provide models for transnational epistemic resilience.
Finally, rehabilitation and reintegration efforts remain vital. Programmes in Indonesia, Singapore, and across Europe show that epistemic change is possible – when personal reflection, community support, and state intervention come together.
Conclusion: The Battle Over How We Know
At its core, the challenge of radicalisation revolves around how people acquire knowledge. Whether in London, Jakarta, Berlin, or Singapore, the same pattern unfolds: grievance is curated online, epistemic closure hardens, and violence becomes plausible.
To confront this, we must look beyond mere content removal or censorship and address the epistemic systems that undergird extremist beliefs. Only by fostering openness, critical inquiry, and mutual understanding – both online and offline – can we reshape the environments in which dangerous ideologies take root.
In an age of AI and algorithmic persuasion, the battle against extremist epistemology is not just about where information is obtained or what people believe, but about how they learn to believe it.
About the Authors
Mohamed Bin Ali is a Senior Fellow in the Inter-Religious Relations in Plural Societies Programme, and Muhammad Haziq Bin Jani is a Senior Analyst in the Indonesia Programme at the S. Rajaratnam School of International Studies (RSIS), Nanyang Technological University (NTU). Sabariah Hussin holds a PhD from the Department of Religion at Rice University in the USA. She is also an alumnus of RSIS. Mohamed Bin Ali is the co-Chairman and Sabariah Hussin a member of the Religious Rehabilitation Group (RRG), Singapore.