Hossain, Muhammad IqbalHossain, Md. IstiakSaha, SwapnilRued, RahnumaRahman, S. M. Sazidur2025-08-212025-08-2120252025-06ID 24341194ID 21301217ID 24241317ID 21201232http://hdl.handle.net/10361/26565Cataloged from PDF version of thesis.Includes bibliographical references (pages 56-57).This thesis is submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in Computer Science and Engineering, 2025.Artificial intelligence (AI), in recent times, is experiencing explosive growth, which comes with improvement and challenges in ethics. In this case, most of the problems raised by deepfake technologies revolve around the issue of changing or producing any sound using sophisticated technology for purposes of mimicking someone’s voice and involves people ranging from ordinary citizens to public figures thus creating a danger to security and privacy. This study mainly concentrates on the important task of recognizing deepfake speech and evaluating the performance of selected speech recognition models in terms of their accuracy and efficiency. Current deepfake speech detection methodologies are explained and their strengths and weaknesses are discussed. The focus of this research work is therefore geared towards designing robust countermeasures against harmful use of deep fake voice technology which enhances trust in electronic communication.57 pagesenBRAC University theses are protected by copyright. They may be viewed from this source for any purpose, but reproduction or distribution in any format is prohibited without written permission.Voice cloningFake audio detectionCNNGRU modelBiLSTM modelArtificial intelligenceArtificial intelligence.Deepfakes.Neural networks (Computer science).Automatic speech recognition.Thesis