Paper Image

Detecting AI-generated text across domains

Published on:

25 March 2024

Primary Category:

Computation and Language

Paper Authors:

Ashok Urlana,

Aditya Saibewar,

Bala Mallikarjunarao Garlapati,

Charaka Vinayak Kumar,

Ajeet Kumar Singh,

Srinivasa Rao Chalamala

Bullets

Key Details

Analyzes statistical, neural and pretrained models for detecting AI text

Comprehensively evaluates approaches on multi-domain mono & multi-lingual data

Achieves 86.9% accuracy on monolingual test set and 83.7% on multilingual

Details experimental setup and error analysis of each technique

Discusses challenges and future perspectives

AI generated summary

Detecting AI-generated text across domains

This paper analyzes various techniques for detecting text generated by large language models across different domains and languages. The methods evaluated include statistical models, neural networks, and pretrained models. Through comprehensive experiments and analysis, the authors achieve over 86% accuracy in detecting AI-generated text. They highlight remaining challenges and considerations for future research.

Answers from this paper

Comments

No comments yet, be the first to start the conversation...

Sign up to comment on this paper

Sign Up