Skip to main navigation Skip to search Skip to main content

Detection of AI-Generated Content Across Modalities: Text, Image and Audio

  • Bonam Sai Sreya
  • , Avasarala Hiranmayi
  • , Shuvishka M. Sajjan
  • , Sanjana Muthukumar
  • , Joanna Grace Fernandez
  • , Soumyalatha Naveen

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

With rapid development in generative models, increase in AI-generated text, images, and videos calls for robust detection across these modalities. This paper aims to develop models which offer a scalable solution capable of being able to differentiate between AI-generated and human-generated content across different modes. The ethical implications of AI-generated content and their significance are also explored. The models are evaluated with publicly available datasets and achieve an accuracy of 99.48% for text, 97.52% for image and 98.46% for audio.

Original languageEnglish
Title of host publicationProceedings of 2025 International Conference on Emerging Technologies in Computing and Communication, ETCC 2025
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798331524760
DOIs
Publication statusPublished - 2025
Event2025 International Conference on Emerging Technologies in Computing and Communication, ETCC 2025 - Bangalore, India
Duration: 26-06-202527-06-2025

Publication series

NameProceedings of 2025 International Conference on Emerging Technologies in Computing and Communication, ETCC 2025

Conference

Conference2025 International Conference on Emerging Technologies in Computing and Communication, ETCC 2025
Country/TerritoryIndia
CityBangalore
Period26-06-2527-06-25

All Science Journal Classification (ASJC) codes

  • Computer Science Applications
  • Control and Systems Engineering
  • Computer Graphics and Computer-Aided Design
  • Hardware and Architecture

Fingerprint

Dive into the research topics of 'Detection of AI-Generated Content Across Modalities: Text, Image and Audio'. Together they form a unique fingerprint.

Cite this