AI-Generated Deepfakes in Cybersecurity Assignment: Evaluating Risks and Detection Strategies

(if the answer to Industry Collaboration is Yes)

(Project Specialisation and Student Specialisation should match)

(Pls. refer at the end of document for the selection of category based on the specialisation)

(Pls. refer at the end of document for the selection of focus/contribution based on the specialisation)

– Ethical Hacking

– Others: AI-Powered Cyber Threat

(Discuss Background, Problem Statement, Methodology, Expected Output/Significance in summary form)

Deepfake technology, powered by artificial intelligence, has become a major cybersecurity concern. Cybercriminals use deepfakes for identity fraud, misinformation campaigns, and social engineering attacks. As deepfake generation tools become more accessible, detecting and mitigating their risks is increasingly challenging.

Problem Statement

Traditional cybersecurity mechanisms struggle to detect AI-generated deepfakes, making organizations and individuals vulnerable to manipulation and fraud. This project aims to evaluate the security risks of deepfake attacks and assess the effectiveness of existing detection techniques.

Methodology

• Literature Review: Analyse existing studies on deepfake technology and its cybersecurity implications.
• Case Study Analysis: Investigate real-world cases of deepfake cyberattacks.
• Security Evaluation: Assess current deepfake detection tools and propose enhancements.
• Recommendations: Suggest strategies to improve deepfake detection and cybersecurity awareness.

Expected Output / Significance

• A comprehensive understanding of deepfake threats in cybersecurity.
• An evaluation of the limitations of current deepfake detection tools.
• A proposed solution in the form of a hybrid detection model that combines AI-based detection with traditional cybersecurity practices for better detection accuracy
• Recommendations on how organizations and individuals can defend against deepfake attacks, reducing their impact on privacy, security, and trust.

(Focused and precise list of statements that can imply the goals to be achieved, Majority of the Project Objectives – using SMART objectives)

2.      To evaluate the security risks posed by AI-generated deepfakes.

3.      To assess existing detection techniques and propose improvements.

(Outcomes are in line with the Project and Student specialisation)

2.      Identification of vulnerabilities in existing security measures.

3.      Recommendations for improving deepfake detection and cybersecurity strategies.

(Focus/Expected Output/ Deliverables with the limits and constraints of the study can be described and implies enough scope for the two-trimester project)

This study focuses on analysing the security risks of AI-generated deepfake attacks in cybersecurity. The project will:

·         Investigate various types of deepfake attacks (e.g., fraud, misinformation, and identity theft) and their impact on cybersecurity.

·         Evaluate the strengths and weaknesses of current detection techniques, identifying gaps in their effectiveness.

·         Propose a solution: A hybrid model for detecting deepfakes by combining AI-based detection tools with traditional cybersecurity measures such as blockchain verification or image authentication (Solution may be adjusted).

Deliverables

1.      Research Report: A comprehensive report on the security risks of deepfakes and their impact on cybersecurity, summarizing key findings from the literature review and case studies.

2.      Evaluation of Detection Tools: An assessment of the effectiveness of existing deepfake detection tools, such as AI-driven models and blockchain-based verification systems.

3.      Solution Proposal: A hybrid detection framework combining AI tools with traditional security measures, designed to enhance deepfake detection accuracy.

4.      Cybersecurity Recommendations: Practical recommendations for organizations to detect and defend against deepfake threats, including best practices for implementing the proposed detection solution

Limitations and Constraints

·         The focus will be on AI-generated deepfakes and cybersecurity implications, not on the technical development of new deepfake models.

·         Due to time constraints (two trimesters), the proposed solution will be tested using simulated datasets or small-scale case studies, rather than real-world deployment.

·         The hybrid detection framework will be theoretical but will provide a concrete roadmap for practical implementation.

(If it is two-students project, subtitles and work distribution must be clearly specified and differentiated for each student)

(Pls. fill up if the number of students is two)

(Pls. fill up if the number of students is two)

(Pls. fill up if the number of students is two)

(Pls. fill up if the number of students is two)

(Student Name, Student ID, Specialisation, Handphone number, E-mail address)

(Leave it blank, if unknown)

Cybersecurity

(Student Name, Student ID, Specialisation, Handphone Number, E-mail address)

(Leave it blank, if unknown)