Detection of AI-Generated Deepfake Voices Using Machine Learning Techniques

Author(s)	Ms. Mehvish Iliyas Patel, Prof. Shakila Siddavatam
Country	India
Abstract	Nowdays AI copies anyone's voice perfectly – sounds completely real. People use this for scams, robbing money, spreading false news. In this project, we built a machine learning system to catch these deepfakes before they cause real damage. We used a CNN model that analyzes tiny audio details like MFCC patterns and spectrograms to spot the fakes. The biggest problem with existing tools? They just say "real" or "fake" but never explain WHY. That's why we added Explainable AI techniques like SHAP and LIME – now users can actually see which parts of the audio looked suspicious. Tested against modern tools like ElevenLabs, we got 92-95% accuracy. Built a complete web app where anyone can upload audio and get instant results with charts showing exactly why the model made its decision. Whether you're a tech expert or regular person, now you can confidently spot AI voice fakes.
Keywords	Deepfake detection, Voice clones, AI scams, Explainable AI, CNN, MFCC, SHAP, LIME, Audio security.
Field	Computer > Artificial Intelligence / Simulation / Virtual Reality
Published In	Volume 8, Issue 1, January-February 2026
Published On	2026-02-27

About IJFMR Fees & Payment Current Issue Publication Archive	Submit Research Paper Track Submission Status Publication Guidelines Publication Ethics Peer Review & Plagiarism	Join as a Reviewer Editors & Reviewers Reviewer Referral Program Get Reviewer Membership Certi.	Website/Journal Policies Usage Policy Content Policies Privacy Policy

Contact Us		+91-9687-828-838	editor@ijfmr.com

International Journal For Multidisciplinary Research