What is the TikHarm Dataset?

The TikHarm dataset is a curated collection of TikTok videos focused on detecting harmful, adult, safe, and suicide-related content. It is designed for research in online safety and harmful content classification.

What is the purpose of the TikHarm Dataset?

The dataset is intended for training and evaluating machine learning models that can detect harmful or inappropriate content on TikTok, ensuring safer online environments for children.

What is the source of the videos?

The videos were collected from TikTok, focusing on content that children are likely to encounter. All videos were carefully reviewed and categorized by experts.

What research applications does this dataset support?

This dataset is suitable for harmful content detection, video classification, online safety research, and developing AI models for child protection and content moderation.

TikHarm Dataset

Q: How is the dataset structured?

The dataset follows a format similar to UCF101 and includes four categories: harmful content, adult content, safe content, and suicide-related content. Each video is manually labeled to ensure accuracy.

Home » Dataset Download » TikHarm Dataset

TikHarm Dataset

Datasets

TikHarm Dataset

File

TikHarm Dataset

Use Case

TikHarm Dataset

Description

Explore the TikHarm dataset, designed to train models for classifying harmful content on TikTok. This curated collection focuses on videos accessible to children, categorized into harmful content, adult content, safe content, and suicide-related material

Description:

The TikHarm dataset is a curated collection of TikTok videos aimed at training models to identify harmful content, specifically focusing on material accessible to children. The dataset is formatted similarly to UCF101 and includes videos categorized into four distinct types: harmful content, adult content, safe content, and suicide-related content.

Download Dataset

Videos were gathered from TikTok with a focus on content that children might encounter. Each video was manually labeled to reflect its category, ensuring that harmful behaviors, inappropriate material, and safe content are clearly identified. This enables the development of models that can accurately classify and filter potentially harmful material for child-friendly environments.

Dataset Structure and Statistics

The dataset is divided into three main subsets:

Training Set: Used for model training and learning patterns
Validation (Dev) Set: Used for tuning and evaluation during training
Test Set: Used for final performance evaluation

Moreover, each subset contains videos of varying durations, ranging from a few seconds to several minutes. Consequently, this ensures diversity and realism in the dataset.

Key Features of the TikHarm Dataset

Manually labeled video dataset for high accuracy
Four distinct content categories for classification
Real-world social media data from TikTok
Suitable for video classification and deep learning models
Designed with child safety and content moderation in mind

Applications and Use Cases

The TikHarm Dataset can be used in a wide range of AI and safety-focused applications. For example:

Content Moderation Systems: Automatically filter harmful or inappropriate videos
Child Safety Platforms: Identify unsafe content for younger audiences
Social Media Monitoring: Improve detection of sensitive or dangerous content
AI Research: Develop advanced video classification and behavior detection models

Conclusion

The TikHarm Dataset is a valuable resource for building AI systems that detect and classify harmful content in videos. With its structured labeling and real-world data, it supports the development of safer digital platforms, particularly for protecting children from inappropriate or dangerous content.

This dataset is sourced from Kaggle.

Contact Us

Let's Discuss your Data collection Requirement With Us

To get a detailed estimation of requirements please reach us.

TikHarm Dataset