Protecting Sensitive Data on USB Drives: An AI-Driven Solution for Malware Detection and Data Privacy
- Version
- Download 3
- File Size 1.02 MB
- File Count 1
- Create Date 7 June 2025
- Last Updated 7 June 2025
Protecting Sensitive Data on USB Drives: An AI-Driven Solution for Malware Detection and Data Privacy
Authors:
Dr. T. AMALRAJ VICTOIRE 1, M. VASUKI 2, J. ANANDHARAJ 3
1 Professor, Department of MCA, Sri Manakula Vinayagar Engineering College, Puducherry-605107, India.
2 Associate Professor, Department of MCA, Sri Manakula Vinayagar Engineering College, Puducherry-605107, India.
3PG Student, Department of MCA, Sri Manakula Vinayagar Engineering College, Puducherry-605107 India.
amalrajvictoire@gmail.com1, dheshna@gmail.com2, anandhraj343@gmail.com3
Abstract: USB drives, external hard drives and memory cards are commonly used for transferring and storing data. These devices are highly portable, easy to use, and can be used virtually anywhere that data is stored. One of the most common modes of covert data theft is by using malware injection attacks, which is when malicious code is injected into a system to secretly collect and/or steal sensitive data. Since these attacks are quite sneaky and can bypass conventional security measures, they can be exploited for unauthorized access of sensitive data. Existing systems tend to focus on either detecting malware or data backup separately. This lack of comprehensive approach does not allow sensitive data to be protected from such attacks. To bridge the gap, this paper presents an integrated security solution that employs deep neural networks (DNN) to detect a malware injection attack, CloudConceal, a secure backup or recovery system, and Data Masking using Tokenization to obfuscate sensitive data stored on USB drives. The DNN model discovers suspicious system activity such as file access and creation of process by analyzing features such as API calls, byte sequences, and metadata of system logs. After finding a malware attack, sensitive data is automatically transported to CloudConceal to create encrypted copies of the data and to recover in case of data loss or compromise of any kind. Besides this, the proposed system obfuscates sensitive information on USB drives in a way that replaces original data with tokens. In return, unauthorized users cannot reach the actual content. This integrated model offers a framework to protect sensitive data from covert data theft, strengthen sensitive data security, and safeguard its availability and integrity.
Keywords: USB security, covert data theft, malware injection attacks, Deep Neural Networks (DNN), AI-driven malware detection, CloudConceal, secure backup and recovery, data masking, tokenization, portable storage protection, data privacy, system activity monitoring, encrypted data, sensitive data protection, cybersecurity.
Download