Requirements Engineering for AI Systems
Deakin University · Completed 2024
Supervised by Dr Chetan Arora
PhD Thesis
Khlood's doctoral research addressed one of the most pressing challenges in modern software engineering: how to specify, elicit, and validate requirements for AI-enabled systems. Unlike conventional software, AI systems exhibit emergent, probabilistic, and often opaque behaviour — characteristics that fundamentally challenge traditional requirements engineering practices developed for deterministic software.
The thesis examined how RE processes, notations, and methodologies need to evolve to accommodate the unique properties of AI systems — including fairness, robustness, explainability, and data quality requirements. The work drew on empirical studies, systematic literature analysis, and case studies to build a grounded understanding of the RE challenges practitioners face and to propose practical methods for addressing them.
Completed at Deakin University, this research has direct impact on how AI governance frameworks — including emerging regulatory standards — can be translated into actionable engineering requirements. Khlood's transition into the Australian Federal Government as an AI Technical Standards Expert is a direct testament to the policy-relevant contribution of this work.