Halil Ibrahim Dursunoglu
Faculty Specialist | Computer Science & Cybersecurity
Western Michigan University
Kalamazoo, Michigan, United States
Research Specialization: Applied Cybersecurity, Artificial Intelligence Security, Secure Software Engineering, Software Vulnerability Detection, Deepfake Detection, and Secure Systems Engineering.
Name Variations in Publications and Professional Records: Halil Ibrahim Dursunoglu; Halil Dursunoglu; Halil I. Dursunoglu; HI Dursunoglu; H.I. Dursunoglu; H. Dursunoglu; Mr. Dursunoglu.
Email: halilibrahim.dursunoglu@wmich.edu
Academic Profiles:
Google Scholar | ORCID | GitHub | ResearchGate
Academic & Research Profile
I am a Faculty Specialist in Computer Science and Cybersecurity at Western Michigan University, specializing in applied cybersecurity and the security of AI-enabled software and digital systems.
My research investigates how software vulnerabilities emerge, persist, and propagate throughout iterative AI-assisted software development. I focus on developing practical methods for vulnerability detection, longitudinal security evaluation, and improving the reliability and trustworthiness of modern computing systems.
My research also addresses AI-generated media security, deepfake detection, semantic and graph-based vulnerability analysis, trustworthy artificial intelligence, and security validation in automated infrastructure.
Drawing on academic research and professional software engineering experience, I seek to bridge the gap between cybersecurity research and real-world implementation through reproducible experimentation, secure systems design, and applied security methodologies.
Research Interests & Expertise
AI-Assisted Software Security: Security of LLM-generated code, iterative software development, vulnerability propagation, security drift, and longitudinal security evaluation.
Software Vulnerability Detection: Static and dynamic program analysis, semantic vulnerability detection, graph-based machine learning, assembly-level security analysis, and automated vulnerability assessment.
Deepfake Detection & Digital Trust: Audio-visual synchronization, AI-generated impersonation, multimedia authenticity, and digital media forensics.
Trustworthy Artificial Intelligence: Evidence-grounded question answering, retrieval-augmented generation, citation verification, and AI reliability.
Secure Systems Engineering: Network security, security policy validation, AI-driven infrastructure automation, configuration security, and resilient computing systems.
Education
Master of Science in Computer Science
Western Michigan University, 2015
Academic & Professional Appointments
Faculty Specialist — Computer Science & Cybersecurity
Western Michigan University | Kalamazoo, Michigan
Responsibilities include undergraduate teaching, curriculum development, applied research, student mentorship, and the development of practical programming and cybersecurity learning activities.
Chief Technology Officer
GrakGrak
Professional experience in technical leadership, software engineering, application development, and systems implementation.
Senior Software Engineer
Dustas
Professional experience in software development, engineering implementation, and software systems.
Peer-Reviewed Journal Publications
1. Hybrid Classification for Complex and Composite Activity Recognition in Multi-User Environments
Halil Ibrahim Dursunoglu et al. (2026)
IEEE Sensors Journal — Accepted for Publication
DOI: https://doi.org/10.1109/JSEN.2026.3741056
2. Audio-Visual Synchronization Analysis for Deepfake Detection: A Comprehensive Review
Halil Ibrahim Dursunoglu (2026)
International Journal of Computer Techniques
Repository DOI: https://doi.org/10.5281/zenodo.19557051
3. Human Activity Monitoring with Wearable Sensors and Hybrid Classifiers
G. Uslu, H. I. Dursunoglu, O. Altun, S. Baydere (2013)
International Journal of Computer Information Systems and Industrial Management
Peer-Reviewed Conference Research
SecurePromptBench: Benchmarking Security Retention in Iterative AI-Assisted Software Development
Halil Ibrahim Dursunoglu (2026)
Sixth Intelligent Cybersecurity Conference (ICSC 2026)
Status: Accepted and Presented; Proceedings Forthcoming.
This research investigates security retention, vulnerability propagation, and longitudinal security evaluation across iterative AI-assisted software development workflows.
Research Preprints & Working Papers
1. SEMA-GUARD: Semantic and Graph-Based Vulnerability Detection in Assembly Code
Halil Ibrahim Dursunoglu and Kaan Sulkalar (2026)
arXiv: https://arxiv.org/abs/2609.17254
This research investigates semantic and graph-based vulnerability detection in low-level software using machine learning and assembly-level program representations.
2. Security Drift in AI-Assisted Software Development: A Framework for Measuring the Security Impact of Iterative Prompt Evolution
Halil Ibrahim Dursunoglu and Kaan Sulkalar (2026)
SSRN: https://doi.org/10.2139/ssrn.7467762
This research examines cumulative security changes and vulnerability propagation throughout iterative AI-assisted software development.
3. Security Weaknesses in LLM-Generated Source Code: An Empirical Vulnerability Analysis of Iterative AI-Assisted Development
Halil Ibrahim Dursunoglu and Kaan Sulkalar (2026)
Zenodo: https://doi.org/10.5281/zenodo.20550245
This study investigates software vulnerabilities associated with LLM-generated source code and repeated code refinement.
4. CiteGuard-RAG: A Validation-Centered AI System for Evidence-Grounded Question Answering
Sumit Barua, Guan Hong, Halil Dursunoglu, Charles Rodgers, Alvis Fong (2026)
arXiv: https://arxiv.org/abs/2609.15830
This collaborative research investigates citation-constrained generation, evidence validation, and improving the reliability of AI-generated responses.
Selected Applied Cybersecurity Research Projects
Security Drift in AI-Assisted Software Development
Research investigating how security vulnerabilities emerge, persist, and accumulate throughout repeated LLM-assisted software development iterations.
The work emphasizes longitudinal security assessment and the development of methods for measuring changes in software security over time.
SecurePromptBench
A research framework for evaluating security retention and vulnerability behavior across iterative AI-assisted software development workflows.
SEMA-GUARD
A semantic and graph-based vulnerability detection framework investigating the use of low-level program representations and machine learning for assembly-level software security.
Deepfake Detection & Audio-Visual Synchronization
Research examining synchronization-based approaches to identifying manipulated digital media, AI-generated impersonation, and threats to multimedia authenticity.
AI-Driven Network Security Drift
Ongoing research investigating how incremental AI-driven network configuration changes can introduce cumulative security risks.
Research interests include longitudinal security requirements, automated configuration validation, continuous security monitoring, and security-preserving remediation.
CiteGuard-RAG
Collaborative research exploring evidence-grounded question answering, citation verification, and validation-centered AI systems.
Editorial, Peer Review & Scholarly Service
Editorial Appointment
Editor — Cyber Serum Journal
Peer Reviewer
- IEEE Access
- IEEE Transactions on Multimedia
- Journal of Computer Science
- Journal of Open Source Software (JOSS)
- Journal of Open Research Software (JORS)
Technical Program Committee
International Conference on Cybersecurity Systems, Technologies and Applications (CSTA 2026)
Technical Program Committee Member.
Conference Presentations & Invited Speaking
BrightTALK Intelligent Network Summit — November 18, 2026 (Upcoming)
When AI Takes Control of the Network: Preventing Security Drift in Automated Infrastructure
Invited presentation on AI-driven network automation, cumulative security risks, continuous validation, and secure infrastructure governance.
BSides Peoria 2026
Security Drift: How AI-Generated Code Gets More Vulnerable Over Time
Conference session: https://sessionize.com/app/speaker/session/1310212
BSides Bloomington 2026
Insert Prompt to Opt Out: Reclaiming Security Control from AI-Generated Code
Presentation on AI-assisted software development security, developer oversight, and vulnerability risks.
Sixth Intelligent Cybersecurity Conference (ICSC 2026)
SecurePromptBench: Benchmarking Security Retention in Iterative AI-Assisted Software Development
Research presentation on longitudinal software security evaluation.
Teaching, Curriculum Development & Mentorship
At Western Michigan University, I contribute to undergraduate computer science and cybersecurity education through teaching, curriculum development, experiential learning, and student mentorship.
My academic activities include:
- Python programming and software development
- Computer organization and assembly language
- Secure programming and cybersecurity fundamentals
- Network security and web application security
- Senior design and applied software engineering projects
- Curriculum redesign and laboratory development
- Honors thesis advising and undergraduate research mentorship
- Cybersecurity outreach and student competitions
My teaching emphasizes practical implementation, secure engineering practices, and preparing students to address real-world cybersecurity challenges.
Media & Expert Contributions
RSA Conference (RSAC)
Expert contribution to Closing the Door on AI Impersonation with Synchronization Scoring, discussing AI impersonation threats and audio-visual synchronization approaches to digital trust.
Cybernews
Expert commentary on Wi-Fi networks, browsing activity, and digital privacy.
Tech Times UK
Expert commentary on artificial intelligence influence, persuasion, and security safeguards.
Video Librarian
Expert commentary on AI-generated media, digital literacy, and the challenges of evaluating synthetic content.
Professional Memberships
- Institute of Electrical and Electronics Engineers (IEEE)
- Association for Computing Machinery (ACM)
- ACM Special Interest Group on Security, Audit and Control (SIGSAC)
Research Significance & Broader Impact
My applied cybersecurity research addresses emerging security challenges associated with the integration of artificial intelligence into software development, digital communications, and automated computing systems.
As AI-assisted software development becomes increasingly widespread, identifying and mitigating vulnerabilities introduced through repeated code generation and refinement is important for improving software assurance and reducing security risks.
My research on security drift and longitudinal vulnerability evaluation seeks to contribute methods for assessing software security throughout development workflows rather than relying exclusively on isolated code assessments.
Research on deepfake detection and AI-generated media security addresses challenges related to digital identity, information authenticity, and AI-enabled impersonation.
My broader research objective is to advance practical cybersecurity methods that improve software reliability, strengthen digital trust, and support the resilience of complex digital infrastructure.
By combining empirical research, secure systems engineering, professional software development experience, and scholarly collaboration, I aim to contribute technologies and methodologies that address important cybersecurity challenges in the United States and beyond.
Academic Profiles & Contact
Halil Ibrahim Dursunoglu
Faculty Specialist, Computer Science & Cybersecurity
Western Michigan University
Kalamazoo, Michigan, United States
Email: halilibrahim.dursunoglu@wmich.edu
Website: https://halildursunoglu.com
Research: https://halildursunoglu.com/research/
Publications: https://halildursunoglu.com/publications/
Google Scholar: https://scholar.google.com/citations?user=v4S2JYEAAAAJ
ORCID: https://orcid.org/0009-0005-7470-9404
GitHub: https://github.com/dursunoglu/
ResearchGate: https://www.researchgate.net/profile/Halil-Dursunoglu