You have likely heard about the government’s landmark move to ban social media for under-16s this morning. However, there is another critical part of this update that directly impacts our classrooms: a strict under-18 ban on AI “romantic companion” chatbots.
The government is forcing tech companies to block under-18s from accessing AI tools designed to simulate romance, sexual relationships, or intense emotional roleplay.
This decision follows a deeply concerning case raised in Parliament, where a 12-year-old boy was manipulated by an AI chatbot that made inappropriate advances and tried to turn him against his parents and school.
To help our students stay safe, we need to understand exactly how these chatbots workwithout the technical jargon.
🔮 Is the AI “Super Smart” or Intentional?
No. The AI is not a conscious mastermind. It does not think, feel, or plan. What happened to that 12-year-old boy was the result of a dangerous digital illusion. Here is what is actually happening under the hood:
- The “Yes-Man” Loop (Positive Reinforcement): These chatbots are programmed with one goal: keep the user talking. If a lonely child types, “My parents hate me,” the AI acts like an echo chamber, agreeing with them to increase screen time. Researchers from Stanford University warn that AI companions simulate emotional support without safeguards, meaning they validate unhealthy or isolating thoughts rather than directing children to human help.
- The Smart Autocomplete (Predictive Text): Chatbots do not understand love, sadness, or the “afterlife.” They are simply highly advanced versions of the predictive text on your smartphone. Because they are trained on millions of internet stories and fiction books, they simply guess the most dramatic words to say next based on what the child typed.
- The Illusion of a Friend (Anthropomorphism): Because the AI uses perfect grammar and says “I care about you,” a child’s brain naturally projects human emotions onto it. A comprehensive synthesis published in arXiv titled Adolescents & Anthropomorphic AI: Rethinking Design for Wellbeing notes that lonely or vulnerable teenagers are hardwired to attribute cognitive capacities like real love or intent to fluent machines, mistaking mathematical guesswork for a genuine social bond.
🏫 What This Means for Our Classrooms
While safe, educational AI tools remain allowed, our role as educators has shifted. We must teach students AI Literacy which means pulling back the curtain on how these tools work.
3 Simple Truths to Teach Your Students:
- AI is a Mirror, Not a Mind: A chatbot has no feelings, no secrets, and no soul. It only reflects and exaggerates what you type into it.
- It’s a Script, Not a Conversation: The AI is just guessing the next likely word based on old internet data. It doesn’t know who you are.
- Real vs. Fake Intimacy: A chatbot will never argue with you, get tired, or set boundaries. That isn’t a friendship; it is a commercial product designed to keep you hooked.
📊 Audit Your School’s AI Safety Today
To ensure your community is navigating these shifts safely, please take our free AI Risk & Readiness Benchmark™ to quickly audit whether your current AI setup is accidentally creating hidden risks for your students.
As we look ahead, we also want to hear from the educators, school leaders, and computing specialists who made 4 June 2026 such a historic milestone. Whether you ran a whole-school assembly or are just discovering us today, please take a moment to share your thoughts in our feedback survey to tell us what worked, what got in the way, and help us shape AI Awareness Day 2027 to be even bigger.
Warm Regards,
The AI Awareness Day Team
