
I read on the LBC website Global that an AI character called Amelia, originally created to help prevent extremism as part of a Home Office-funded video game, has been co-opted by far-right groups online.
What started as a counter-extremism project intended to help young people avoid radicalisation has now been turned into a viral symbol on social media — with far-right users turning Amelia into memes and videos spreading nationalist and anti-immigrant messages.
But what’s happened since is astonishing. Instead of staying in that educational context, far-right groups online have taken Amelia and turned her into a symbol — creating memes, AI videos and social posts that promote nationalist, xenophobic messages.
In these clips you might see her marching through London or the House of Commons, declaring love for Britain and spreading harmful tropes about immigrants and religious groups.
The trend spread incredibly fast — going from a few hundred posts a day to thousands — and has even spawned an Amelia-themed cryptocurrency that’s been shared widely on social media.
So a project that was meant to help steer young people away from harmful ideologies has instead been repurposed as a viral far-right icon — something that shows us just how powerful and unpredictable AI and online culture can be.
So what can parents do?
Stories like this can sound frightening. But the answer isn’t panic. And it isn’t pretending this world doesn’t exist.
It’s connection, conversation, and building children’s inner compass.
Here are some gentle, practical ways parents can respond.
1. Don’t start with fear. Start with curiosity.
Ask what your child is seeing. What AI characters, memes, videos or influencers are popping up in their world?
Not:
“Have you seen anything dangerous?”
But:
“Show me what’s trending.”
“What do kids your age find interesting online right now?”
“What do you think about it?”
Children open up when they don’t feel interrogated.

2. Talk about influence, not just safety.
Help children understand that a lot of online content isn’t neutral.
AI characters, memes and viral videos are often designed to:
• grab attention • stir emotion • create identity • push ideas
A powerful question to teach is:
“Who made this — and what do they want me to feel or believe?”
That one question builds lifelong digital resilience.
3. Explain that not everything online is what it seems.
Children are growing up in a world where:
• people aren’t always real • images aren’t always true • and AI can be shaped by anyone
They need to know that characters, avatars and influencers can be hijacked, remixed and weaponised.
Not to scare them — but to sharpen their awareness.
4. Strengthen identity offline.
Extremist content doesn’t usually hook children with politics.
It hooks them with:
• belonging • certainty • anger • being “special” or “chosen”
Strong family connection, real conversations, community, hobbies, and feeling valued offline make children far less vulnerable online.
A grounded child is much harder to manipulate.
5. Supervise, but don’t secretly spy.
Yes, use filters and boundaries.
But always pair them with explanation.
“I’m not watching because I don’t trust you.
I’m guiding because the online world isn’t built with children’s wellbeing at heart.”
Protection works best when it’s transparent and relational.
6. Make home a safe place to question ideas.
Children need somewhere they can say:
“I saw this.”
“I didn’t understand that.”
“It made me feel uncomfortable.”
“It kind of made sense.”
If we shut conversations down, the internet becomes their teacher.
If we stay open, we stay influential.
7. Focus on long-term skills, not one-off scares.
The goal isn’t to protect children from one story.
It’s to raise young people who can:
• question • reflect • resist emotional manipulation • and think for themselves
Because childhood today isn’t about avoiding the digital world.
It’s about learning to live wisely inside it.

Quick Parent Scripts: When Stories Like “Amelia” Appear
1. The curiosity opener
“Can you show me what kinds of AI characters or memes people your age are sharing at the moment? I’m really interested in what’s popping up in your world.”
(This keeps the door open.)
2. The influence conversation
“Not everything online is created just to entertain. A lot of things are designed to shape how people think or feel. When you see something powerful, it’s always worth asking: ‘Who made this, and what might they want me to believe?’”
3. The AI awareness script
“Some of the characters and videos you see online aren’t real people at all. They can be created or changed by anyone, anywhere, for any reason. That’s why it’s okay — and clever — to question what you’re looking at.”
4. The reassurance script
“If something online ever feels confusing, extreme, uncomfortable or even convincing, you don’t have to figure it out on your own. You can always bring it to me. My job isn’t to tell you off. It’s to help you make sense of things.”
5. The boundary script
“I do set limits around screens — not because I don’t trust you — but because the online world isn’t built around children’s wellbeing. It’s built around attention. And you’re more important than algorithms.”

6. The critical thinking builder
“When something spreads really fast, that doesn’t mean it’s true. It usually means it’s emotional. Strong feelings travel faster than facts.”
7. The resilience script
“You don’t need the internet to tell you who you are or where you belong. You already belong — here, with us, and in the real world around you.”
8. The closing anchor
“The online world will keep changing. My role is to help you grow a strong inner compass — so you can enjoy technology without being pulled around by it.”
Final thought
This story about “Amelia” shows us something important.
Technology will keep changing.
AI will keep evolving.
And online spaces will keep being contested.
But the most powerful protection children have has never changed.
A trusted adult.
An open door for conversation.
And the skills to think, not just consume.
Because childhood matters. And so does how we protect it.
You can read the LBC article here
