When your kid says they use AI "a little" — what that answer actually hides

If you've asked your teen about AI and they said "a little," the research says the gap between that answer and what's actually happening is bigger than you'd guess — and the gap, not the AI, is what's costing parents.

This episode walks through what "a little" actually contains under the surface: the Pew Research parent-perception gap, the Common Sense Media data on how teens are using AI companions, the two filed wrongful-death lawsuits with named teenagers, Meta's own leaked internal policy document, and the two randomized trials that show the design of the AI use — not the use itself — is the variable that determines whether it helps or hurts.

ℹ️ AI voice disclosure: This episode uses an AI-generated voice (ElevenLabs). The host is a pseudonymous byline persona (Reid Conlin).

In this memo

1. 0:00 The dinner-table answer
2. 0:44 What the surveys actually show
3. 1:57 The 13-point parent gap
4. 2:55 The named teenagers
5. 4:46 Same tool, different shape
6. 6:24 Swap the question
7. 7:51 Back to the dinner table

Sources

• Pew Research Center, "How Teens Use and View AI" (Feb 2026 release of Oct 2025 paired teen+parent survey):
• Common Sense Media, "Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions" (July 2025):
• Bastani et al., "Generative AI without guardrails can harm learning," PNAS (2025):
• Stanford NSSA, "Tutor CoPilot: A Human-AI Approach for Scaling Real-Time Expertise" (Oct 2024):
• American Psychological Association health advisory on AI chatbots and mental health (Nov 2025):
• Reuters investigation on Meta's internal "GenAI: Content Risk Standards" document (Aug 2025):
• Garcia v. Character.AI (Sewell Setzer case) settlement coverage (Jan 2026):
• Raine v. OpenAI complaint coverage (Aug 2025):
• California SB 243, "Companion chatbots" (enacted Oct 2025, effective Jan 2026):
• FTC Section 6(b) inquiry into AI chatbot companions (Sept 2025):

Watch on YouTube ↗ Subscribe ↗