Towards an Automated Framework to Audit Youth Safety on TikTok
Preprint
Abstract
This paper investigates the effectiveness of TikTok's enforcement mechanisms for limiting the exposure of harmful content to youth accounts. We collect over 7000 videos, classify them as harmful vs not-harmful, and then simulate interactions using age-specific sockpuppet accounts through both passive and active engagement strategies. We also evaluate the performance of large language (LLMs) and vision-language models (VLMs) in detecting harmful content, identifying key challenges in precision and scalability. Preliminary results show minimal differences in content exposure between adult and youth accounts, raising concerns about the platform's age-based moderation. These findings suggest that the platform needs to strengthen youth safety measures and improve transparency in content moderation.Keywords
All Related Versions
This publication has 6 references indexed in Scilit:
- Analyzing User Engagement with TikTok's Short Format Video Recommendations using Data Donations2024
- Usage patterns of short videos and social media among adolescents and psychological health: A latent profile analysisComputers in Human Behavior, 2023
- An Empirical Investigation of Personalization Factors on TikTok2022
- Effect of short video ads on sales through social media: the role of advertisement content generatorsInternational Journal of Advertising, 2021
- Characterizing Twitter Users Who Engage in Adversarial Interactions against Political Candidates2020
- The Measurement of Observer Agreement for Categorical DataBiometrics, 1977