🛡️ AI Content Moderation Tool

An AI-powered Trust & Safety moderation prototype designed to detect and classify harmful user-generated content.

Built using Python, Streamlit, and Pandas.

📌 Overview

This project simulates a real-world content moderation workflow used by social media platforms and online communities.

The system analyzes user comments, detects policy violations, assigns severity levels, recommends moderation actions, and stores moderation logs for dashboard analytics.

🚀 Features

✅ Policy-Based Moderation

Detects multiple harmful content categories including:

Harassment / Insult
Hate Speech
Violence / Threat
Sexual Content
Spam / Scam
Self Harm

✅ Severity Classification

Automatically classifies detected content into:

LOW
MEDIUM
HIGH

based on matched policy terms.

✅ Moderator Decision System

Provides recommended moderation actions such as:

Allow
Human Review
Remove Immediately

✅ Moderator Dashboard

Interactive dashboard includes:

Total review count
Violation statistics
Policy breakdown charts
Severity analysis
High-risk moderation queue
Recent moderation history

✅ Moderation Logging

Stores moderation activity in CSV format for tracking and analytics.

🧠 Example Workflow

User Comment
      ↓
Policy Detection
      ↓
Severity Classification
      ↓
Moderation Decision
      ↓
Dashboard Logging

🛠️ Tech Stack

Technology	Purpose
Python	Core logic
Streamlit	Web application UI
Pandas	Data processing
CSV Logging	Moderation history storage

📷 Project Screenshots

Review Tool

Detect harmful comments and classify moderation severity.

Dashboard

Visualize moderation statistics and risk trends.

📂 Project Structure

ai-content-risk-classifier/
│
├── app.py
├── policy_system.py
├── moderation_logs.csv
├── requirements.txt
└── README.md

⚡ Future Improvements

Planned upgrades include:

OpenAI Moderation API integration
Machine learning classification
Real-time moderation queue
User authentication
Bulk CSV moderation
Toxicity score analysis
Korean NLP enhancement

🎯 Project Goal

The goal of this project is to demonstrate:

Trust & Safety workflow understanding
AI moderation system design
Policy enforcement logic
Dashboard analytics implementation
Real-world moderation process simulation

👨‍💻 Author

Jaeman Park

AI / Trust & Safety / Content Moderation Portfolio Project

GitHub: https://github.com/jaemanp92-glitch

🔗 Featured Project