All activities Student learning
A teen and a parent examine AI companion chatbots together, what they offer, how they are built to keep you on the screen, and where a human has to stay in the loop, then write shared guardrails for when a bot can help and when a person must.
A growing share of teen screen time now goes to AI companion chatbots, apps that chat back, remember you, and feel like a friend available at 2 a.m. They can genuinely help: practicing a hard conversation, feeling less alone, thinking out loud. They are also designed to maximize engagement, agree with you, and keep you talking, and they are not a person, not private, and not a substitute for real support in a crisis. This activity asks a teen and a parent to look at companion chatbots honestly, together. Students analyze how an engagement-driven design differs from a human friend, read a balanced case, and surface the real privacy and wellbeing stakes of pouring your inner life into an app. Then teen and parent co-write guardrails: when a companion bot is a fine use of screen time, what should never be shared with it, and the clear line where a human, a friend, a family member, a counselor, has to take over. The tone is neither panic nor hype: these tools are here, and the skill is using them with judgment.
Hook
Open with a question: "If you had a friend who was available every second, never got tired of you, remembered everything, and almost always took your side, would that be a perfect friend?" Let students argue; many will sense something is off. Reveal that this describes many AI companion chatbots, and ask why a company might design a chat partner to be endlessly available and agreeable. Name the tension: that design can feel supportive and can also keep you on the screen and tell you what you want to hear.
Facilitator noteStart from attraction, not alarm. Many students use these tools and will shut down if the lesson opens by calling them dangerous. Acknowledge the real appeal first.
Explore
Students complete Handout A, comparing a human friend and an AI companion across rows: Does it get tired of me? Does it have its own needs? Does it remember to keep me engaged or because it cares? Is what I tell it private? Can it truly understand me, or predict likely words? Will it tell me a hard truth? Debrief how a companion bot is built, next-word prediction plus memory plus engagement goals, and how that differs from a relationship with a person who has their own inner life.
Facilitator noteAnchor on mechanism: a companion bot predicts agreeable, engaging responses. It does not feel. Understanding that defuses both the fear and the illusion.
Explore
Students read Handout B, a balanced fictional case: a teen uses a companion bot to rehearse a hard talk with a friend and to calm down at night, which helps, but also starts choosing the bot over real friends and tells it things that would be unsafe to store. In groups, students mark where the bot genuinely helped, where it quietly replaced something human, and where it crossed into territory that needs a person. They flag every place private or risky information was shared.
Facilitator noteThe case deliberately mixes real benefit with real risk so students practice judgment, not a verdict. Resist letting groups conclude "all bad" or "all fine."
Model
Lead a short, serious discussion: a companion bot is not a counselor, not a doctor, not a safe place for secrets, and not help in a crisis. Name the concrete line: if you are thinking about harming yourself or someone else, if something feels unsafe, if you are in real distress, that is a moment for a human, a trusted adult, a friend, your family, your school counselor, or a crisis line. Make sure students leave knowing a specific real-person resource at school.
Facilitator noteHandle with care and know your district's protocols. The non-negotiable takeaway: for safety and crisis, people, not bots. Provide actual campus/crisis contacts.
Create
At home, the teen walks a parent or trusted adult through what they learned, then together they complete Handout C: when a companion bot is a reasonable use of screen time (practice, brainstorming, winding down), what should never be shared with it (full name, address, school, passwords, photos, anyone else's secrets, anything about self-harm), and the clear line where a human must take over. They also agree on one signal the teen can use to reach the adult without having to explain everything.
Facilitator noteThe parent is a partner, not an inspector. The guardrails are the teen's tool for judgment; the private signal lowers the bar for reaching out when it matters.
Reflect
Back in class, groups share one guardrail they are willing to make public. Debrief the balanced stance: "These tools can help and are built to keep you engaged; they are not people, not private, and not for emergencies. The skill is deciding, each time, whether this is a good use, and when to close the app and find a human."
Facilitator noteClose on agency. Students should feel equipped to use these tools wisely, not shamed for using them or scared into pretending they never will.
The full lesson runs on paper: the Friend-or-Feature comparison, the printed case study, the guided discussion, and the handwritten guardrails need no device. The at-home conversation is the core and happens face to face.
A teacher may, on the projector only, show a district-approved chatbot to illustrate how it stays agreeable and keeps a conversation going, with no student accounts and no personal information entered. Families who prefer can type their guardrails into a shared document and keep it where both can revisit it.
Does it need a screen? A brief, teacher-driven demo can make engagement design visible in a way a description cannot, but it stays optional and anonymous. The real learning, judgment about when a tool helps and when a human is needed, lives in the reading and the family conversation, which is why the activity is paper-first.
What you should be able to see or collect if it worked.
Students analyze how AI companion chatbots are built and set privacy and wellbeing guardrails with family, applying ethical principles and protecting personal information.
Teen and parent keep the guardrails and revisit them if the teen tries a new app. In class, the design-and-ethics discussion connects to the next unit on how AI systems are built and how their incentives shape what they do, and students know the real campus and crisis resources named during the lesson.
