# When Youth Enter the Algorithmic Wild: Discovering and Understanding Potentially Harmful Teen Videos on Douyin and Kwai

arXiv: https://arxiv.org/abs/2605.23598 · HTML: https://mlsanity.com/paper/2605.23598/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Safety alignment, refusal and guardrails, one of 12 directions in LLM safety and security.
- Builds on 19 of its 44 references that are in the map.
- First posted on arXiv 22 May 2026.

## Placed in

- LLM safety and security / [Safety alignment, refusal and guardrails](https://mlsanity.com/safety/safety-alignment-refusal-and-guardrails/)

## Builds on (most cited first, 19 of 19)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [Qwen-VL: A Versatile Vision-Language Model for Understanding, Localization, Text Reading, and Beyond](https://mlsanity.com/paper/2308.12966/) (2023)
- [ToxicChat: Unveiling Hidden Challenges of Toxicity Detection in Real-World User-AI Conversation](https://mlsanity.com/paper/2310.17389/) (2023)
- [In-Context Learning with Long-Context Models: An In-Depth Exploration](https://mlsanity.com/paper/2405.00200/) (2024)
- [Fake News Detection on Social Media using Geometric Deep Learning](https://mlsanity.com/paper/1902.06673/) (2019)
- [A BERT-Based Transfer Learning Approach for Hate Speech Detection in Online Social Media](https://mlsanity.com/paper/1910.12574/) (2019)
- [MOMENTA: A Multimodal Framework for Detecting Harmful Memes and Their Targets](https://mlsanity.com/paper/2109.05184/) (2021)
- [Many-Shot In-Context Learning in Multimodal Foundation Models](https://mlsanity.com/paper/2405.09798/) (2024)
- [FireRedASR: Open-Source Industrial-Grade Mandarin Speech Recognition Models from Encoder-Decoder to LLM Integration](https://mlsanity.com/paper/2501.14350/) (2025)
- [HateMM: A Multi-Modal Dataset for Hate Video Classification](https://mlsanity.com/paper/2305.03915/) (2023)
- [Your fairness may vary: Pretrained language model fairness in toxic text classification](https://mlsanity.com/paper/2108.01250/) (2021)
- [Convolutional Neural Networks for Toxic Comment Classification](https://mlsanity.com/paper/1802.09957/) (2018)
- [Towards Explainable Harmful Meme Detection through Multimodal Debate between Large Language Models](https://mlsanity.com/paper/2401.13298/) (2024)
- [An Empirical Investigation of Personalization Factors on TikTok](https://mlsanity.com/paper/2201.12271/) (2022)
- [Harmful YouTube Video Detection: A Taxonomy of Online Harm and MLLMs as Alternative Annotators](https://mlsanity.com/paper/2411.05854/) (2024)
- [Supporting Human Raters with the Detection of Harmful Content using Large Language Models](https://mlsanity.com/paper/2406.12800/) (2024)
- [Dynamics of Algorithmic Content Amplification on TikTok](https://mlsanity.com/paper/2503.20231/) (2025)
- [Detecting and Understanding the Promotion of Illicit Goods and Services on Twitter](https://mlsanity.com/paper/2404.07797/) (2024)
- [MTikGuard System: A Transformer-Based Multimodal System for Child-Safe Content Moderation on TikTok](https://mlsanity.com/paper/2511.17955/) (2025)

## Abstract (arXiv)

Short-video platforms like Douyin and Kwai have become central to adolescent digital life, but they also risk exposing teens to algorithmically amplified harmful content. Despite its societal importance, the scale, mechanisms, and real-world impact of this exposure remain poorly understood. Measuring it is challenging: recommendation feeds are personalized black boxes, harmful content employs sophisticated evasion tactics, and naive crawlers fail to replicate authentic teen behavior. To bridge this gap, we propose PHTV-Scout, the first large-scale, behaviorally grounded measurement framework for Potentially Harmful Teen Videos (PHTVs). We integrate an offline survey of 683 adolescents with a tri-module online pipeline: (1) PHTV Hunter simulates teen accounts to collect recommendation feeds; (2) PHTV Arbiter, a LoRA-finetuned multimodal classifier, detects PHTVs with 94.29% accuracy and 96.41% precision; and (3) PHTV Analyzer performs fine-grained categorization and impact assessment. Over six months, we analyzed 186,727 videos and 51,287 comments, uncovering a troubling 6.11% PHTV prevalence--dominated by Child Sexual Exploitation Imagery (53.2%)--and revealing that harmful content thrives through covert interactions (e.g., grooming comments, self-disclosure) and active evasion (semantic camouflage, noise injection). Crucially, while Youth Mode blocks 100% of PHTVs, its low adoption (30-41%) leaves most teens unprotected. We further show that exposure is driven not by user identity but by regulation, platform algorithms, and even passive browsing, exposing the fragility of adolescent information environments. Our findings call for a paradigm shift from reactive takedowns to proactive, human-centered safeguards.