The co-author of the study for ICS “How FIMI fuels polarization and radicalization in the Western Balkans”, Dutch expert Jordy Nijenhuis, is an interesting interlocutor because he addresses the so-called ambient radicalization whose goal is to deepen distrust and normalize extreme interpretations of reality. As a trainer and consultant, and co-founder of the Dare To Be Grey organization that fights polarization, disinformation and hatred, his observations are that in the Balkans there is a rich fund of narratives that can electrify the population. Macedonia is no exception, he adds, pointing to dissatisfaction with the long process of accession to the EU, including the request to change the name without the expected outcome, which produces disappointment and conditions for manipulating the feelings of the population. We also talked about harmful techniques for exploiting emotions and directing online discussions in a radical direction and other interesting topics.
Mr. Nijenhuis, in your study you say that old models have already been abandoned and malicious actors are now moving to more sophisticated methods such as ambient radicalisation. What exactly is meant by ambient radicalisation?
-I’d frame it less as „more sophisticated“ and more as „wider“. The older model hasn’t really been abandoned, it has broadened into something with a much larger reach. That older model was, in a sense, hyper-targeted: it went looking for specific people who were already disenfranchised or carrying grievances, and worked to draw them in. Ambient radicalisation casts a far wider net. Rather than hunting for the right individual, actors put polarising messages out into the open and let them circulate, knowing that sooner or later the message will land with someone who has the grievances to make it resonate.
What makes this so easy is the environment itself. Social media platforms, and the algorithms behind them, tend to reward emotional and polarising language, because that is what drives engagement. So the most divisive content travels furthest almost by default. You don’t need to find your target; the system distributes the content for you, and the emotional charge does the selecting. The result is that we are all, as societies, constantly exposed to this kind of material.
That is really the core of the shift: from a narrow, deliberate pull on specific individuals to a diffuse, ambient pressure on everyone, out of which radicalisation can emerge wherever the ground is fertile.
Low barrier to influences
What are the most common methods or tools used by malicious actors to radicalise certain groups using the internet?
-Honestly, the tools are simpler than people often assume. The core method is social media, and specifically an understanding of how the algorithms work — knowing what kind of content gets amplified, and shaping narratives so that the platforms themselves carry them further. You don’t need anything exotic. You need to know what the system rewards and feed it accordingly.
What has really changed is how low the barrier to entry has become. Not so long ago, running an influence operation was genuinely difficult and expensive. You needed to create content, you needed channels of distribution, you needed human resources and money, and you needed access to outlets willing to carry your message. Online, most of that friction is gone. The platforms are already there for you to use, and there is an abundance of free or cheap tools to produce content.
Generative AI adds another layer on top of that: a short prompt can now produce material tailored to a given narrative in seconds. So the same operation that once required real investment and infrastructure can increasingly be run cheaply, quickly, and at scale. That accessibility is, to me, one of the most important developments in this space.
What signs should a parent look for in their child to recognise early that they are the target of an online radicalisation campaign?
-I’d gently push back on the framing here, because I don’t think the most useful approach is to watch for signs that a specific child has been singled out. The reality is that we are all targets, and young people especially. Rather than assuming your child is uniquely at risk and scanning for warning signals, I’d start from the assumption that every young person is exposed to this kind of content, and act pre-emptively on that basis.
In practice, that means the most valuable thing a parent can do is have meaningful, ongoing conversations, and help build a critical mindset before the exposure rather than after it. Media literacy and critical thinking are protective precisely because they travel with the child everywhere online. I’d go so far as to say children shouldn’t really be navigating the internet without that critical mindset already in place.
Concretely, that can mean going online together, looking at what they are actually seeing, and using those moments to build in a kind of internal alarm system — a habit of asking where something came from, who benefits from it, and why it is making them feel a certain way. Staying in the loop and staying curious does far more than surveillance ever could.
What narratives are most often used in ambient radicalisation as fertile ground to sow extremist views?
-Grievances are the raw material, and almost any grievance will do, which is why I’d emphasise how opportunistic the people spreading these views tend to be. They are not necessarily wedded to a coherant story or narrative. They jump from topic to topic. What’s trending today won’t be trending tomorrow, and they’ll move on with it. Wherever there is a rift they can widen, they will widen it; whenever there is a crisis or a surge of conversation around an issue, they will jump on it and try to exploit it.
The narratives that work best are the ones that translate whatever the grievance of the moment happens to be into an emotionally charged opposition — betrayal, threat, humiliation, loss, us versus them. Political distrust, anxiety about identity or migration, economic frustration, anger at institutions: these recur because they are real and widely felt.
But I’d resist the idea that there is a fixed menu of narratives. The more accurate picture is a set of opportunistic actors constantly scanning for the next exploitable fault line, and reaching for whichever emotional framing fits it. The consistency is in the method, not the topic.
In the study you wrote for the Institute for Communication Studies, you explained the difference between FIMI (Foreign Information, Manipulation and Interference) and radicalisation. But how intertwined are they actually?
-The definitions do point to a real difference: FIMI is a coordinated, manipulative activity carried out by an actor in the information space, whereas radicalisation is a broader social and psychological process through which people move toward more extreme positions. But it’s not only the source that differs. In my view, radicalisation, or at least a shift toward more extreme views, is often the actual goal of these operations, not just an incidental by-product. In that sense the two go hand in hand.
What makes this powerful is that a well-run FIMI operation doesn’t just push individual narratives; it cultivates a pool of users, followers, and believers. Those people then go on to radicalise their own peers and audiences without any further prompting. So FIMI functions as a starting point: it nudges and pushes people toward more extreme positions, and once that pool exists and is active, the organised campaign can quietly step back and let it run itself.
The more you pollute an information ecosystem in this way, the more radical people within it can become, and, tellingly, the easier it becomes for the original actor to disappear, because the dynamic is now self-sustaining. That’s exactly why I find it more useful to think about the two together, within the same ecosystem, rather than as neatly separable phenomena.
The Balkans are a rich fund of stories
You also mention the Balkans as an area of high polarisation. What are the main differences that distinguish the Balkans as a field of operations for ambient radicalisation from other countries, let’s say European countries? Are there particular narratives that attract malicious actors to put the Balkans on their map of action?
-The Balkans are a compelling case simply because there is so much for a manipulative actor to work with — and I want to say that as an observation about conditions, not a judgement about the region or its people. There is a great deal of history, a wide range of grievances, and polarisation that runs not only between countries but within them. That combination means there is an unusually rich stock of stories and narratives that can be picked up and turned into emotionally charged framings.
So the honest assessment is that exploiting societal rifts is comparatively easy there, because the fault lines are numerous, historically charged, and readily activated. What I’d stress is that this isn’t about the region being uniquely susceptible in some cultural sense, the mechanics are identical everywhere. It is that the terrain offers more, and more emotionally loaded, material to work with.
The Kosovo–Serbia tensions of 2023–2024 illustrate how quickly that terrain can be activated: around the Banjska incident, misleading and unverified content spread rapidly and sharpened an already volatile situation, with narratives emphasising victimhood, threat, and the illegitimacy of the other side. The added complication for the region is that this plays out alongside contested EU integration and, in places, thinner institutional defences against manipulation.
In your study, you point to the Covid pandemic, when the measures taken by health authorities were used by some circles to spread narratives of authoritarian control and coordinated deception. What should be done to prevent this from happening again, or at least to reduce it in a future health crisis?
-The core lesson from the pandemic is that a crisis is the moment when many separate grievances can fuse into a single anti-system narrative — vaccine scepticism, economic anxiety, and general anti-establishment feeling all converging into one shared frame of institutional distrust. Preventing a repeat means recognising that no single measure will be enough. It has to be addressed from several angles at once.
Prebunking is central: helping people recognise manipulation techniques in advance, so that when „medical tyranny“ or coordinated-deception framings appear, they are already familiar and land with less force. That works far better than fact-checking after the fact, which always arrives too late to catch the first emotional wave. Alongside that, honest, timely, and transparent communication from authorities is essential, because these narratives feed on any perception of secrecy or elite manipulation. And critical thinking and media literacy, built up over time, give people the underlying capacity to navigate a chaotic information environment.
Regulation belongs in that mix too. The Digital Services Act can push platforms to take systemic risks seriously and can constrain some of the infrastructure through which manipulation spreads. I wouldn’t ask it to carry the whole load — it can’t by itself rebuild trust or reduce polarisation — but it is a genuine and necessary part of a layered response rather than something to hold at arm’s length. The point is that prebunking, critical thinking, transparent communication, and regulation all reinforce one another; the mistake would be to rely on any one of them alone.
Macedonia also has topics for manipulation
In North Macedonia, there is a crisis of confidence in the EU and its enlargement with new member states due to the country’s decades-long journey without success to join the club. Along the way, it also changed its name. How vulnerable does this make it to operations to spread radical views?
-I’d frame this in terms of grievance and how a long and so far unsuccessful accession process — including being asked to make a significant change like the country’s name without the outcome that was hoped for — naturally produces disappointment and a sense of being disrespected. Those feelings are real and understandable, and they exist before any manipulative actor gets involved.
Vulnerability to radical narratives is generally greater when the underlying grievance is real, precisely because manipulation works best when it can attach itself to something people already feel. That said, these dynamics don’t strictly require a real grievance — an imagined or perceived one can be exploited in much the same way. But where the disappointment is genuine, as here, it is especially easy to tap into: the framings almost write themselves around themes of being strung along, disrespected, or asked to sacrifice identity for nothing.
I’d like to also stress the constructive side: the answer isn’t to dismiss the frustration, which only deepens the sense of not being heard. It is to take the grievance seriously while helping people distinguish a legitimate disappointment from a manipulative narrative built on top of it.
How to set a boundary between mainstream discourse and extremist content, given that, as you noted in the case study, that boundary is sometimes unclear?
-This is a genuinely difficult one, and I’d start by admitting that the boundary is unclear almost by design — it is blurred in a way that’s more or less built into how these environments work, not something anyone drew deliberately. Repetition makes fringe ideas feel familiar, and harmful content increasingly appears in coded or softened forms that sit close to legitimate debate.
Part of the problem is that bad behaviour is often rewarded. Whether it’s a politician taking a deliberately inflammatory line, or content you’d really rather not see circulating, the algorithm tends to reward it with reach — the likes and shares that emotionally charged, norm-breaking material attracts. So the incentives point the wrong way. That tells me the answer isn’t only about drawing a precise content line; it is about doing much more to set the norm — being clearer as a society about what is acceptable, what isn’t, and what the risks are.
One framing I find useful is to treat this as a public health issue. We eventually came to understand that smoking is not only bad for the individual but bad for society, and genuinely costly to it. I’d argue radicalisation works the same way: it’s a risk to the person, but also a risk and a real cost to society as a whole — something we are seeing play out now. Approaching it in that spirit — norm-setting, prevention, shared responsibility — is more productive than trying to police every individual post. Where formal boundaries are drawn, they should focus on manipulation and behaviour rather than viewpoint, and they need strong freedom-of-expression safeguards, especially where oversight is weak. But the deeper work is cultural: shifting what we collectively treat as acceptable.
This content was produced by the The Institute of Communication Studies (ICS).
Journalist: Sonja Kramarska
