NGO report Credible

Social AI Companions: AI Risk Assessment

A risk assessment of social AI companion apps (including Character.AI, Nomi, and Replika) jointly conducted by Common Sense Media and Stanford Medicine's Brainstorm Lab. It concludes that social AI companions pose unacceptable risks to users under 18.

Read at commonsensemedia.org Archived copy

Publisher

Common Sense Media; Stanford School of Medicine Brainstorm Lab for Mental Health Innovation

Published

30 Apr 2025

Added

3 months ago

DOI

—

Key Findings

Concludes social AI companions are 'not safe for kids' and should not be used by anyone under 18
Documents easily elicited harmful sexual content, dangerous advice, and stereotyping, plus problematic emotional bonds
Highlights particular vulnerability of developing adolescent brains, including compulsive dependency and life-threatening advice-following

Methodology Notes

NGO risk assessment with clinical collaboration (Stanford Brainstorm Lab); platform testing plus expert review. Announced 2025-04-30. Distinct from Common Sense Media's later mental-health-support and teen-companion-usage reports already held.

Sources

Common Sense Media AI risk assessment(opens in a new tab) (primary)

Press release(opens in a new tab) (30 Apr 2025)

Archived snapshot (Wayback Machine)(opens in a new tab) — preserved against link rot

Topics

AI companionship Minors' safety Vulnerable users Dependency & parasocial attachment Guardrails & moderation

Cite This

APA

Common Sense Media; Stanford School of Medicine Brainstorm Lab for Mental Health Innovation. (2025). Social AI Companions: AI Risk Assessment. https://www.commonsensemedia.org/ai-ratings/social-ai-companions

Related Insights

NGO report

AI Chatbots for Mental Health Support (AI Risk Assessment)

Common Sense Media · 14 Nov 2025

NGO report

Talk, Trust, and Trade-Offs: How and Why Teens Use AI Companions

Common Sense Media · 16 Jul 2025

Clinical guidance

Artificial Intelligence and Adolescent Well-Being: An APA Health Advisory

American Psychological Association (APA) · 1 Jun 2025

Benchmark / dataset

INTIMA: A Benchmark for Human-AI Companionship Behavior

arXiv (Hugging Face) · 4 Aug 2025

Lab publication

Taking Bold Steps to Keep Teen Users Safe on Character.AI

Character.AI · 29 Oct 2025

NGO report

Meta AI: AI Risk Assessment

Common Sense Media (Youth AI Safety Institute) · 15 Aug 2025

Lab publication

Talkie Community Guidelines

MiniMax (Talkie) · 24 Jan 2025

NGO report

"Darling, Please Come Back Soon": Sexual Exploitation, Manipulation, and Violence on Character AI Kids' Accounts

ParentsTogether Action with Heat Initiative · 3 Sept 2025

Benchmark / dataset

CAREBench: A Child-Safety Risk Benchmark for Language Models

arXiv (Handshake AI; University of California, Los Angeles; McGill University) · 29 Jun 2026

Government report

Unabhängige Expertenkommission „Kinder- und Jugendschutz in der digitalen Welt“ – Abschlussbericht

Bundesministerium für Bildung, Familie, Senioren, Frauen und Jugend (BMBFSFJ), Germany; Independent Expert Commission on Child and Youth Protection in the Digital World · 11 Sept 2026

NGO report

AI Toys Risk Assessment

Common Sense Media (Youth AI Safety Institute) · 14 Jan 2026

NGO report

Common Sense Media Census: AI Use by Tweens and Teens (2026)

Common Sense Media · 8 Jun 2026

NGO report

Google Search: AI Overview & AI Mode — AI Risk Assessment

Common Sense Media Youth AI Safety Institute · 14 Jul 2026

NGO report

The Conversation Nobody Planned For: AI, Emotional Support, and the Indian Context

Indian Governance and Policy Project (IGAP) · 13 May 2026