Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{203610,
        author = {Aiswariya A and Abisha J and Adithya S Kumar and Dhivya V M},
        title = {Deepfake Audio Detection with Neural Networks Using AI and Audio Features},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {no},
        pages = {188-193},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=203610},
        abstract = {AI-generated deepfake audio poses serious risks to security, misinformation, and identity fraud, creating a need for reliable detection systems. This project develops a deepfake audio detection framework using spectral and temporal voice features, voiceprint analysis, and AI-based classification models. The system combines Convolutional Neural Networks (CNN) for feature extraction and Bi-LSTM networks for temporal pattern analysis to accurately distinguish synthetic speech from real human voices.
A real-time filtering and probabilistic scoring mechanism further improves verification and detection reliability. Trained on diverse datasets with multiple speech synthesis techniques, the model analyzes pitch, tone, and spectral inconsistencies to provide a scalable and effective solution against AI-driven audio manipulation.
Existing deepfake detection techniques often struggle to adapt to evolving speech synthesis methods and may fail to identify subtle inconsistencies in audio signals. Therefore, this project focuses on developing a robust and adaptive system using Convolutional Neural Networks (CNN) and Bi-LSTM networks to analyze both spectral and temporal characteristics of speech. In addition, the integration of real-time filtering and probabilistic scoring mechanisms aims to provide accurate, fast, and reliable verification suitable for real-world applications such as media authentication, voice-based security systems, and online communication platforms.
Attackers can create convincing fake audio recordings to impersonate public figures, company executives, or ordinary individuals for malicious purposes. Traditional authentication and security systems are not designed to detect such advanced manipulations, highlighting the need for intelligent and automated detection frameworks.},
        keywords = {.},
        month = {May},
        }