How Radicalization Happens Online - and What It Looks Like in Public Posts
Quick answer: Radicalization online is usually a slow tilt, not a sudden flip - a feed drifts from ordinary to angry to a worldview where one group is to blame for everything. You cannot see the private groups and recommendation loops doing most of the work, but you can read the public trail: the reposts, the hardening us-versus-them frame, the coded language creeping in. Read across time, keep the actual posts, and treat a clean scan as "nothing public stood out," not proof anyone is fine. This is a personal read of public posts, not a background check.
Radicalization is rarely a lightning strike. It looks more like a slow tilt - a feed that used to be memes and match highlights, then a few angry reposts, then a worldview where one group is behind everything that ever went wrong. The person at the center almost never feels the slope under their feet. The people around them sometimes do, because the change shows up in what gets posted.
Most of the machinery runs where you cannot see it: private groups, closed chats, a recommendation engine nudging one more video, then another. But the public trail still matters, because it is the part you can actually read. Here is how radicalization happens online in plain terms, and, more usefully, what the drift looks like in someone's public posts before it hardens into something they will defend.
The slow tilt, not the switch
People want a villain moment - the day a normal person decided to believe something monstrous. It almost never works like that. What happens is closer to erosion. A real grievance (a lost job, a breakup, plain loneliness) meets a corner of the internet that offers a tidy explanation and a crowd that welcomes you for accepting it. The explanation always ends the same way: your problem has a cause, the cause is those people, and everyone too polite to say so is part of it.
The pull is social as much as ideological. The group is warm to newcomers and brutal to doubters, so agreeing feels like belonging and questioning feels like exile. Each step is small enough to seem reasonable from the last one. That is why the person rarely notices - and why the clearest evidence is not any single post but the direction of travel across many.
How radicalization happens online, step by step
Strip away the specifics and the shape repeats across movements. First a grievance gets named and aimed - vague frustration becomes a specific enemy. Then comes community: an account, a group, a comment section that rewards the new framing with attention. Next the world splits cleanly in two, us and them, and the "them" bucket quietly widens to include former friends. Finally the language shifts. Ordinary words get swapped for in-group shorthand, and violence that was once unthinkable gets reframed as defense.
None of those stages announce themselves. But each leaves fingerprints on a public feed - the moment someone starts reposting a harder account, the first time a joke has a real target, the reply where a whole group gets written off. You are not watching a switch flip. You are watching a slope, and slopes are visible if you look at more than one point on them.
Worried about someone's feed - a partner, a family member, or your own? ACCOUNTability! reads thousands of a person's public posts across X, TikTok, Instagram, Facebook and LinkedIn and flags extremist, hateful and conspiracy content, with the actual post as receipts, so you are judging what is really there. €15.
Read the public postsReading the drift in public posts
You do not need training to see a slope; you need to stop staring at one post. None of this is rare, either - about 56% of Americans say they have experienced online hate or harassment in their lifetime (ADL, 2024), which is the same water a slow tilt takes hold in. Here is a way to read a public feed honestly, without turning one bad day into a verdict:
- Read across time, not one post, and compare what the account sounded like a year ago with what it sounds like now.
- Watch the reposts and replies, because who a person amplifies often shifts before their own words do.
- Look for a hardening us-versus-them frame - the same group blamed for unrelated problems, over and over.
- Notice new vocabulary: coded numbers, slogans and 'just asking questions' framings borrowed from a specific crowd.
- Track the tone, not just the topic, since grievance curdling into contempt is the shift that matters.
- Keep the actual posts with dates, so you are reacting to a pattern you can see, not a feeling you cannot.
Run that read and one thing gets easier: telling a genuinely worrying drift apart from an ordinary person having an ordinary bad week. A single furious post after some piece of news is noise. The same enemy surfacing in unrelated threads, month after month, alongside steadily harsher company - that is the pattern worth taking seriously. A scan just does this reading at scale, pulling the flagged posts into one place so you look at evidence instead of chasing a hunch through years of timeline.
The honest limits of a scan
Be straight about what a public read can and cannot do. It sees public accounts only - the private groups and closed chats where a lot of the real pull happens are simply out of reach, so quiet flags are not an all-clear. It only has something to work with if the person actually posts; a barely-used account gives you little. And it is AI flagging content with the receipts attached, which means context can trip it: sarcasm, reclaimed language, or an argument quoted to mock it can all read wrong at a glance, which is exactly why the actual post sits next to every flag for you to judge.
One more line worth keeping straight. A read of public posts is personal due diligence, not a background check or consumer report, and it plays no part in any regulated employment, tenancy or credit decision - use a licensed provider for those. A clean result means nothing in the person's public posts stood out. Not that they are safe, not that they are fine, and never that they have been vetted. It is a starting point for a conversation, not the end of one.
Key takeaways
- Radicalization online is a gradual drift, not a single moment, so the pattern over time matters more than any one post.
- The public tell often shows up in what a person reposts and replies to before it shows up in their own words.
- A hardening us-versus-them frame - one group blamed for unrelated problems - is the clearest sign the tilt is underway.
- Much of the process happens in private spaces a public-post scan cannot see, so an absence of flags proves nothing.
- A scan reads public posts only and flags content with the receipts; a clean result means nothing public stood out, not that the person is safe.
Common questions
Can you really tell how radicalization happens online from public posts?
You can see the public part of it, which is often enough to notice. The private groups, closed chats and recommendation loops that do a lot of the work stay hidden, but the drift usually leaks into what a person posts in the open: harder reposts, a fixed enemy, new coded language. Read across time rather than fixating on one post, and you can watch the tilt taking shape. What you cannot do is prove intent from a feed - you are reading signals, not verdicts.
What are the early signs someone is being pulled in?
The tone shifts before the topics do. Grievance starts curdling into contempt, the same group gets blamed for unrelated problems, and the accounts they amplify get more extreme than anything they would say themselves yet. New vocabulary shows up - coded numbers, slogans and 'just asking questions' framings borrowed from a specific crowd. One angry post is nothing; a steady drift in one direction over months is the thing to notice.
Can ACCOUNTability! check someone's public posts for extremist content?
Yes. ACCOUNTability! reads thousands of a person's public posts across X, TikTok, Instagram, Facebook and LinkedIn and flags extremist, hateful and conspiracy content, showing you the actual post behind each flag so you decide what it means. It costs fifteen euros, it reads public posts only, and it is a personal read, not a background check or consumer report. A clean result means nothing public stood out, not that the person is safe or verified.
See the drift before it hardens
If a feed you care about has been tilting, ACCOUNTability! reads thousands of that person's public posts across X, TikTok, Instagram, Facebook and LinkedIn and flags extremist content, hate speech, transphobia and conspiracy stuff - each flag shows the actual post so you can judge it yourself. There are tools that do this for companies; as far as we know, nothing built for regular people. €15 a scan, no sales call.
Run a scan