Seeking Help, Facing Harm: Auditing TikTok's Mental Health Recommendations

2026-04-16

AI Summary

This study investigates whether TikTok’s recommendation system can differentiate between users expressing psychological distress and those actively seeking help, and whether it accordingly delivers appropriate mental health content. By deploying 30 controlled accounts guided by large language model (LLM)-based interactive agents, the authors simulated varied initial search and engagement strategies over seven days to audit content appearing on the “For You” page. Integrating LLM-driven agents with a controlled experimental design, this work offers the first quantitative assessment of the platform’s sensitivity to user intent and its content safety boundaries. Results indicate that user engagement significantly shapes recommendations: active interaction increased mental health–related content to 45% of the feed, while passive browsing still yielded 11–20% exposure. Although help-seeking queries elevated supportive content, harmful material related to suicide or self-harm persisted at low frequencies.

Technology Category

• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval
• Humans and AI: Interaction Techniques and Devices
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction

Application Category

• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Web Mining and Content Analysis: Content-based information diffusion
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

Recommender systems on social media increasingly mediate how users encounter mental health content, yet it remains unclear whether they distinguish help-seeking from distress expression. We conduct a controlled 7-day audit of TikTok's "For You" page using 30 fresh accounts and LLM-guided agents that vary initial search framing (distress- vs. help-initiated) and interaction strategy (engaged, avoidant, passive). Across 8,727 recommended videos, interaction behavior dominates exposure outcomes: engagement rapidly saturates feeds with mental health content (~45% of daily recommendations), while avoidance and passive viewing reduce but do not eliminate exposure (~11-20%). Search framing mainly shifts composition rather than volume--help-initiated searches yield more potentially supportive material, yet potentially harmful content persists at low but non-zero levels, including content in the Suicide/Self-Harm category. These findings suggest limited sensitivity to user intent signals in TikTok's recommendations and motivate context-aware safeguards for sensitive topics.

Problem

• recommender systems
• mental health
• help-seeking
• distress expression
• content moderation

Innovation

• algorithmic auditing
• LLM-guided agents
• mental health recommendations
• context-aware safeguards
• social media recommender systems