# AI-Generated Phishing Detection Dataset

**18k emails**: 11k safe + 7k phishing (traditional + 5 synthetic AI-generated from LLM prompts).

## Columns
- `Unnamed: 0`: ID
- `Email Text`: Full email body
- `Email Type`: "Safe Email" / "Phishing Email"
- `label`: 0=safe, 1=phish

## Usage
```python
from datasets import load_dataset
ds = load_dataset("premsaidhulipala/ai-phishing-dataset")