Title: Leveraging AI for enhanced deepfake detection! Developing framework for designing safe ecosystem to safeguard digital authenticity

Authors: Pooja Gupta; Vijay Kumar Jain; Shrish Singh; Srabanti Maji; Shipra Agarwal

Addresses: School of Computing, DIT University, Dehradun, India ' Department of Management Studies (SoLAM), DIT University, Mussoorie Diversion Road, Makka Wala, Dehradun, Uttarakhand-248009, India ' Department of Management Studies (SoLAM), DIT University, Mussoorie Diversion Road, Makka Wala, Dehradun, Uttarakhand-248009, India ' School of Computing, DIT University, Dehradun, India ' Department of Commerce, Graphic Era Deemed to be University, Dehradun (UK), India

Abstract: The deepfake problem, defined as the creation and spread of modified movies or images that appear legitimate but are actually fake, has far-reaching effects. Deepfakes threaten to destroy the basic fabric of truth and confidence in society, needing immediate and extensive efforts to counteract their development and distribution. Addressing the deepfake issue is critical to maintaining information integrity, defending privacy rights, and restoring trust in our digital environment, therefore, the current study is a modest attempt to develop a framework for deepfake detection using AI. The variables having high impact and relationships with deepfake detection have been identified. Total 15 variables were identified based on literature and experts' opinion. The fuzzy DEMATLE-AHP has been applied on select variables to prioritise them in order of their effectiveness and find causality among them. Seven drivers were classified as effect whereas as eight drivers were identified as causes for deepfake detection. The analysis shows that blockchain technology (DF12), penalties (DF14) and watermarking and digital signature were ranked as the most significant drivers for deepfake detection.

Keywords: digital forensics; blockchain; synthetic media; machine learning; voice cloning; DEMATLE.

DOI: 10.1504/IJICS.2025.148859

International Journal of Information and Computer Security, 2025 Vol.28 No.2, pp.145 - 180

Received: 24 Jul 2024
Accepted: 16 Jan 2025
Published online: 29 Sep 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article