YouTube Script AI Detector

Fine-tuned RoBERTa model to detect AI-generated YouTube scripts.

Model Description

This model was fine-tuned on 203 labeled YouTube script examples (102 AI, 101 human) to distinguish between AI-generated and human-written content specifically in the context of YouTube videos.

Training Data

  • AI examples: Structured explainers, tech reviews, gaming content, documentaries, horror narration
  • Human examples: Vlogs, personal stories, casual content, streams, tangential storytelling

Performance

  • Accuracy: 100%
  • F1 Score: 1.0
  • Precision: 1.0
  • Recall: 1.0

Usage

from transformers import pipeline

classifier = pipeline("text-classification", model="ekrembalkis/youtube-ai-detector")
result = classifier("Your YouTube script text here")
print(result)
# [{'label': 'AI' or 'HUMAN', 'score': 0.99...}]

Limitations

  • Trained specifically on YouTube script style content
  • May not generalize well to other content types
  • Dataset size is relatively small (203 examples)
  • May need retraining as AI writing styles evolve

Citation

If you use this model, please cite:

@misc{youtube-ai-detector,
  author = {Emity},
  title = {YouTube Script AI Detector},
  year = {2024},
  publisher = {HuggingFace},
  url = {https://huggingface.co/ekrembalkis/youtube-ai-detector}
}
Downloads last month
24
Safetensors
Model size
0.1B params
Tensor type
F32
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for ekrembalkis/youtube-ai-detector

Quantizations
1 model

Space using ekrembalkis/youtube-ai-detector 1