Abstract:
The rapid advancement of artificial intelligence has enabled the creation of hyper-realistic synthetic media, commonly known as deepfakes. While this technology has potential applications in entertainment, education, and creative industries, it also poses serious threats to privacy, security, and the integrity of digital information. Deepfakes can be weaponized for disinformation campaigns, identity theft, and cybercrimes, making their detection a pressing research challenge. This paper explores AI and machine learning-based approaches for detecting deepfakes in images, audio, and video content. Specifically, it highlights the use of convolutional neural networks (CNNs), recurrent neural networks (RNNs), transformer- based architectures, and multimodal learning techniques to identify subtle inconsistencies in visual frames, facial expressions, and audio- visual synchronization. Furthermore, the study discusses benchmark datasets, evaluation metrics, and real-world applications of detection systems. Finally, challenges such as dataset limitations, adversarial attacks, and the evolving sophistication of generative models are addressed, alongside recommendations for developing robust, ethical, and scalable deepfake detection solutions.
Page(s):
32-32
DOI:
DOI not available
Published:
Journal: 4th International Conference of Sciences “Revamped Scientific Outlook of 21st Century, 2025” , November 12,2025, Volume: 1, Issue: 1, Year: 2025
Keywords:
Convolutional neural networks
,
Recurrent Neural Networks
,
HyperRealistic Synthetic Media
,
Deepfakes