Deepfake Detection Using Hybrid Models

Farooq Sunar Mahammad, Gajula Geetha, Gopireddy Thanusha, Atkur Manasa, Daruri Harika, Kolakani Jahnavi

2025

Abstract

Deepfake technology is doing real-world damage, aggravating concerns over online fraud, identity theft, and the distribution of misinformation. Powered by artificial intelligence, it can generate disturbingly lifelike fake videos, pictures and even voices that are hard to tell apart from real material. This study examines the detection of deepfakes, which identifies the potential of machine learning in supporting advanced AI models, including transformers, recurrent neural networks (RNNs), and convolutional neural networks (CNNs). We consider whether various machine learning techniques are effective at spotting deepfakes by analysing facial movement, image patterns and audio cues to highlight small inconsistencies. But identifying deepfakes isn’t straightforward problems include restricted training data sets, constantly changing manipulation methods and attacks meant to deceive detection systems. To solve these problems, we also study "XAI (Explainable AI)" which makes AI decisions transparent and much more interpretable. This research aims to develop more robust, scalable, and AI-driven approaches that enhance the detection accuracy of deepfake technology, protect against the loss of digital authenticity, and safeguard against potential abuses. We seek to develop tools that can perform real-time detection across the pro- to anti- spectrum of the content that circulates in social media, news sites and other online environments through multi-modal analysis and large datasets. We also discuss the ethical and legal implications of deepfake technology, highlighting the importance of regulations and collaboration among researchers, policymakers, and tech companies. As deepfake technology improves, it’s important to stay ahead of detection technology as well. This research aims to connect state-of-the-art AI developments with real digital world use cases to protect and provide a safer and more reliable world for all of us.

Paper Citation

in Harvard Style

Mahammad F., Geetha G., Thanusha G., Manasa A., Harika D. and Jahnavi K. (2025). Deepfake Detection Using Hybrid Models. In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25; ISBN 978-989-758-777-1, SciTePress, pages 911-922. DOI: 10.5220/0013875700004919

in Bibtex Style

@conference{icrdicct`2525,
author={Farooq Mahammad and Gajula Geetha and Gopireddy Thanusha and Atkur Manasa and Daruri Harika and Kolakani Jahnavi},
title={Deepfake Detection Using Hybrid Models},
booktitle={Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25},
year={2025},
pages={911-922},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013875700004919},
isbn={978-989-758-777-1},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies - Volume 1: ICRDICCT`25
TI - Deepfake Detection Using Hybrid Models
SN - 978-989-758-777-1
AU - Mahammad F.
AU - Geetha G.
AU - Thanusha G.
AU - Manasa A.
AU - Harika D.
AU - Jahnavi K.
PY - 2025
SP - 911
EP - 922
DO - 10.5220/0013875700004919
PB - SciTePress