A Lightweight Approach to Detection of AI-Generated Texts Using Stylometric Features

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Aityan, Sergey K., Claster, William, Emani, Karthik Sai, Rais, Sohni, Tran, Thy
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915717757009920
author Aityan, Sergey K.
Claster, William
Emani, Karthik Sai
Rais, Sohni
Tran, Thy
author_facet Aityan, Sergey K.
Claster, William
Emani, Karthik Sai
Rais, Sohni
Tran, Thy
contents A growing number of AI-generated texts raise serious concerns. Most existing approaches to AI-generated text detection rely on fine-tuning large transformer models or building ensembles, which are computationally expensive and often provide limited generalization across domains. Existing lightweight alternatives achieved significantly lower accuracy on large datasets. We introduce NEULIF, a lightweight approach that achieves best performance in the lightweight detector class, that does not require extensive computational power and provides high detection accuracy. In our approach, a text is first decomposed into stylometric and readability features which are then used for classification by a compact Convolutional Neural Network (CNN) or Random Forest (RF). Evaluated and tested on the Kaggle AI vs. Human corpus, our models achieve 97% accuracy (~ 0.95 F1) for CNN and 95% accuracy (~ 0.94 F1) for the Random Forest, demonstrating high precision and recall, with ROC-AUC scores of 99.5% and 95%, respectively. The CNN (~ 25 MB) and Random Forest (~ 10.6 MB) models are orders of magnitude smaller than transformer-based ensembles and can be run efficiently on standard CPU devices, without sacrificing accuracy. This study also highlights the potential of such models for broader applications across languages, domains, and streaming contexts, showing that simplicity, when guided by structural insights, can rival complexity in AI-generated content detection.
format Preprint
id arxiv_https___arxiv_org_abs_2511_21744
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle A Lightweight Approach to Detection of AI-Generated Texts Using Stylometric Features
Aityan, Sergey K.
Claster, William
Emani, Karthik Sai
Rais, Sohni
Tran, Thy
Computation and Language
Artificial Intelligence
I.2.7; I.5.1
A growing number of AI-generated texts raise serious concerns. Most existing approaches to AI-generated text detection rely on fine-tuning large transformer models or building ensembles, which are computationally expensive and often provide limited generalization across domains. Existing lightweight alternatives achieved significantly lower accuracy on large datasets. We introduce NEULIF, a lightweight approach that achieves best performance in the lightweight detector class, that does not require extensive computational power and provides high detection accuracy. In our approach, a text is first decomposed into stylometric and readability features which are then used for classification by a compact Convolutional Neural Network (CNN) or Random Forest (RF). Evaluated and tested on the Kaggle AI vs. Human corpus, our models achieve 97% accuracy (~ 0.95 F1) for CNN and 95% accuracy (~ 0.94 F1) for the Random Forest, demonstrating high precision and recall, with ROC-AUC scores of 99.5% and 95%, respectively. The CNN (~ 25 MB) and Random Forest (~ 10.6 MB) models are orders of magnitude smaller than transformer-based ensembles and can be run efficiently on standard CPU devices, without sacrificing accuracy. This study also highlights the potential of such models for broader applications across languages, domains, and streaming contexts, showing that simplicity, when guided by structural insights, can rival complexity in AI-generated content detection.
title A Lightweight Approach to Detection of AI-Generated Texts Using Stylometric Features
topic Computation and Language
Artificial Intelligence
I.2.7; I.5.1
url https://arxiv.org/abs/2511.21744