Everyday Safety

Signs Someone Is Being Radicalized Online

By the ACCOUNTability! Team · 7 min read

Quick answer: Radicalization rarely shows up as one dramatic post. The signs someone is being radicalized online are usually a pattern: a slow shift toward dehumanizing, us-versus-them language, new extremist accounts they follow and reshare, and posts that keep getting angrier and more absolute. Most of that is readable from public posts alone. But context matters. One dark joke is not a verdict, a private account gives you little to go on, and a clean read means nothing public stood out, not that everything is fine.

Nobody announces it. There is no single post where a person you know declares they have gone down a hole. It happens in the drift: a reshare here, a new account they start following there, a comment section they slowly move into and stop leaving. By the time the change is obvious at the dinner table, the trail online is usually months long.

So if you are worried about a brother, an old friend, an adult in your family, or someone you are getting close to, their public posts are where that drift shows first. This walks through the signs someone is being radicalized online, where to actually look, and how to tell a real pattern from one edgy joke you would be wrong to hang someone on.

How radicalization actually shows up

Forget the movie version, the one big unhinged rant that gives it all away. Real radicalization is gradual and it hides in plain sight, spread across small choices: what a person laughs at, who they defend, which stranger's video they decide to boost to their own followers. Each piece looks minor. Stacked over months, they point somewhere.

And this is not a fringe worry you can assume will never reach anyone you know. About 56% of Americans say they have experienced online hate or harassment in their lifetime (ADL, 2024). The pipelines that pull people toward that kind of content are busy, and they are good at making the next step feel like a small one.

The signs someone is being radicalized online

No single item on this list means much by itself. The weight is in how many show up, and whether they keep sharpening.

Language starts dehumanizing a group. A whole category of people becomes an infestation, a disease, a plot. When individuals stop being individuals and turn into a faceless "them," that is the grammar of extremism doing its work.
Everything becomes us versus them. The world flattens into an in-group that gets it and an out-group that is either stupid or evil. Nuance vanishes. Disagreement gets read as betrayal.
New vocabulary appears. Coded slurs, movement slang, numbers and symbols that mean nothing to you but everything inside a particular corner. If a word showed up in their posts this year and you have to look it up, look it up.
The feed they amplify changes. Watch the reshares and follows, not just the original posts. People often boost the harder stuff before they will say it themselves.
The tone escalates. Compare this month to last spring. Angrier, more absolute, more certain that violence or exclusion is the only answer left. Escalation is the tell that separates a phase from a slide.

You do not have to read a year of someone's timeline by hand. ACCOUNTability! reads thousands of a person's public posts across X, TikTok, Instagram and Facebook and flags extremist, hateful and conspiracy content, each flag showing the actual post so you can judge it yourself. €15.

Read their public posts

Where to look

The signal is spread across platforms, and it rarely all sits in one place. Someone can keep a tidy, bland Facebook for relatives while a second account elsewhere runs hot. So look wide: original posts, yes, but also replies, quote-posts, the accounts they follow, and the videos they reshare. Following and resharing is where people try out a position before they own it out loud.

Stay on the public side of the line. This is about what a person has chosen to broadcast, not their private messages, their locked accounts, or anything you would have to trick your way into. If it is not public, it is not yours to read, and it is not part of this.

A checklist for reading the pattern

Start with what is public: their posts, reshares, replies and the accounts they follow across X, TikTok, Instagram and Facebook, not private messages you cannot and should not read.
Read for a pattern over time rather than a single post, because one angry joke is noise and the same theme every week is signal.
Note shifts in vocabulary such as dehumanizing language, hard us-versus-them framing, coded slurs or movement slang that were not there a year ago.
Watch who they amplify: whose videos they reshare, whose accounts they suddenly follow, and which comment sections they now live in.
Look at escalation, checking whether the posts are getting angrier and more absolute or holding steady where they were.
Keep the receipts and judge each actual post in context before you say anything, since sarcasm and reclaimed language can read worse than they are meant.

What a scan can and can't tell you

Be straight with yourself about the edges of this. A read of public posts sees what a person has made public and nothing else. A locked account, a private group chat, a deleted post history, or a burner you have never found all stay dark. Someone quiet online can be in real trouble and leave almost no trace; someone loud can be all posturing.

It is AI flagging content and handing you the receipts, not a verdict. Sarcasm, a reclaimed slur, a screenshot shared to mock it, a lyric quoted out of context, all of these can trip a flag that a human would clear in two seconds, which is exactly why every flag comes with the actual post attached. Your job is to read it in context. And a clean result does not certify anyone. It means nothing in their public posts stood out, not that the person is safe, and not that there is nothing to worry about somewhere you cannot see. Treat what you find as a reason to pay attention and, if it is bad, to have a real conversation, not as a file to weaponize.

Key takeaways

Radicalization is a pattern, not a single post: look for repetition and escalation over weeks, not one bad take on a bad day.
The tells sit in the language (dehumanizing, us-versus-them, coded slurs or slang), in who a person amplifies, and in a feed that keeps getting angrier.
Most of it is visible in public posts across X, TikTok, Instagram and Facebook, including the accounts they follow and reshare, without touching anything private.
Context still matters: sarcasm and reclaimed language can read worse than they are, which is why you judge the actual post, not just the flag.
This reads public posts only and is not a background check; a clean result means nothing public stood out, not that someone is safe.

Common questions

What are the signs someone is being radicalized online?

The clearest signs someone is being radicalized online are a pattern rather than a single post: language that dehumanizes a whole group, hard us-versus-them framing, coded slurs or movement slang, and a feed that increasingly amplifies extremist accounts and gets angrier over time. A one-off edgy joke is not proof of anything. What matters is repetition and escalation across weeks, in public posts you can actually see.

Can you tell if someone is radicalized without their private messages?

Often, yes. Radicalization usually leaves a public trail: the accounts a person follows, the videos they reshare, the comment sections they live in, and the way their own posts drift. You do not need private messages to see a pattern forming. What public posts cannot show you is intent, or anything set to private, so treat what you find as a reason for a real conversation, not a diagnosis.

Is checking someone's public posts for this a background check?

No. Reading a person's public posts for extremist or hateful patterns is personal due diligence on things they already made public, not a background check or consumer report, and it plays no part in any employment, tenancy or credit decision. ACCOUNTability! reads thousands of public posts across X, TikTok, Instagram and Facebook and flags extremist, hateful and conspiracy content for fifteen euros, showing you the actual post. A clean result means nothing public stood out, not that a person is safe.

Read the posts, not the panic

If you are worried about someone, ACCOUNTability! reads thousands of their public posts across X, TikTok, Instagram and Facebook 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

or see a real example report

Full disclosure: ACCOUNTability! is our own tool - this is the company blog. It reads public accounts only, it is a personal check of public posts and not a background check, and a clean result means nothing public stood out, not that someone is safe.