What question did this study set out to answer?

The survey aims to assess and compare various AI-driven methods for detecting deepfake videos.

April 11, 2026Open Access

Deepfake Video Identification and detection method using Artificial Intelligence

Key Points

The survey aims to assess and compare various AI-driven methods for detecting deepfake videos.
Comparative analysis of AI methods for deepfake detection
Evaluation of Generative Adversarial Networks (GANs)
Examination of Convolutional Neural Networks (CNNs)
Analysis of Long Short-Term Memory (LSTM) networks
Assessment of hybrid CNN and Jaya Algorithm approach
Hybrid CNN with Jaya Algorithm achieved 99.3% accuracy, the highest among methods tested
Hybrid LSTM-CNN method displayed low accuracy at 82%
CNN-Jaya method demonstrated the highest precision rate
Recall rate of CNN-Jaya method was robust but slightly lower than other approaches
Identification of limitations and challenges in current deepfake detection methods

Abstract

This survey explores the landscape of deepfake video identification and detection methods, leveraging the capabilities of artificial intelligence (AI). As the proliferation of deepfake technology poses significant challenges to the veracity of digital content, the need for robust and efficient detection mechanisms becomes paramount. The survey delves into various AI-driven approaches employed for discerning deepfake videos from authentic content, including Generative Adversarial Networks (GANs), Convolutional Neural Networks (CNNs), and recurrent architectures like Long Short-Term Memory (LSTM) networks. Based on the comparative analysis, the hybrid LSTM-CNN method is ineffective in detecting deepfakes, with a low accuracy rate of 82%. However, the hybrid method combining CNN with the Jaya Algorithm significantly improves accuracy to 99.3%, outperforming all other methods in detecting deepfake videos. This method also achieves the highest precision rate, though its recall rate, while robust, is slightly lower than some other techniques. The study addresses the limitations and challenges associated with these methods, including the constant evolution of deepfake techniques and the ethical considerations surrounding privacy and consent. By providing a comprehensive overview of the current state of deepfake video identification, this survey aims to contribute valuable insights for researchers, practitioners, and policymakers navigating the intricate landscape of AI-powered deepfake detection..

Mark Helpful

Bookmark

Relay

View Full Paper