Adolescents & Anthropomorphic AI: Rethinking Design for Wellbeing An Evidence-Informed Synthesis for Youth Wellbeing and Safety

TL;DR

This paper explores how anthropomorphic design in conversational AI can support adolescent wellbeing and safety by balancing risk mitigation with autonomy and skill development, informed by developmental science and policy insights.

Contribution

It introduces a behavioral framework and global policy dialogue to guide anthropomorphic AI design for adolescent wellbeing, bridging science and industry practices.

Findings

Anthropomorphism can be used as a design lever for risk mitigation.

Non-negotiable guardrails are essential for safe AI interactions with adolescents.

Design strategies support adolescents' autonomy and skill development.

Abstract

Conversational AI has become part of adolescents' everyday lives. This report asks: what does AI owe adolescents when it can speak to them like a social partner? The synthesis bridges the gap between developmental science and industry practice through consultations, a behavioral framework, and global policy dialogue. It identies non- negotiable guardrails and highlights the role of anthropomorphism as a design lever for risk mitigation, ensuring systems support adolescents' autonomy and skill development.