Home Research Responsible & Safe AI
Research Theme · 02
← Back to Research

Responsible &
Safe AI Systems.

As AI systems become embedded in consequential decisions, the question shifts from 'does it work?' to 'does it behave as intended, remain fair over time, and fail safely?' This theme addresses specification, monitoring, and governance of AI systems.

Theme illustration

Theme Overview

What we study

AI systems are rarely built from scratch — they are trained, deployed, and updated in messy real-world environments. This theme focuses on the requirements and engineering methods needed to keep AI systems trustworthy: before deployment (specification and fairness requirements), during operation (runtime monitoring), and as they evolve (agentic AI guardrails and governance frameworks).

Key research questions

  • ·How do we specify fairness and ethical requirements for AI-based applications in a way that is verifiable?
  • ·How can runtime monitoring detect when an ML system violates its behavioural requirements?
  • ·What guardrails are needed for LLM-based agentic systems making autonomous decisions?
  • ·How do existing regulations (EU AI Act, GDPR) translate into engineering requirements?
  • ·How do users perceive and report ethical concerns in AI-powered apps?

Related Publications

Selected Papers in this Theme.

[ACM TOSEM · A*] 2025 Monitoring Requirements for ML-Based Systems
+

Research on runtime monitoring approaches for ML systems — ensuring deployed models continue to satisfy their behavioural requirements as data distributions shift. Led by PhD student Hira Naveed.

Google Scholar ↗

[ICSME'25] 2025 Ethical Requirements in LLM-Based Healthcare Applications
+

Empirical study identifying ethical concerns raised by users of LLM-powered healthcare apps — informing a requirements framework for responsible clinical AI. Led by PhD student Yutan Huang.

Google Scholar ↗

[JSS · A] 2025 Fairness Requirements in AI-Based Software Systems
+

Systematic approach to eliciting, specifying, and verifying fairness requirements for AI systems — covering protected attributes, fairness metrics, and their translation into testable requirements.

Google Scholar ↗

[Applied Soft Computing] 2024 Guardrails for LLM-Powered Agentic Systems
+

Analysis of safety guardrail mechanisms for autonomous LLM-based systems — covering bounded autonomy, scope specification, and fallback behaviour requirements.

Google Scholar ↗

[IREB RE-Magazine] 2024 Ethics of Using LLMs in Requirements Engineering
+

Practitioner-facing analysis of the ethical dimensions of integrating LLMs into requirements engineering workflows — covering bias, accountability, and transparency. Invited article for the International Requirements Engineering Board publication.

Google Scholar ↗

People

Researchers in this Theme.

Associate Supervisor
Safeguarding the Development and Usage of Agentic AI Applications in the Enterprise Context
Monash University
Associate Supervisor
Model-Driven Engineering for Runtime Monitoring of Human-Centric Requirements in ML Systems
Monash University
Associate Supervisor
Agentic AI-Driven Modular Orchestration for Trustworthy Personalised Wellness
Monash University

Related PhD Research

RY Rui Yang HN Hira Naveed SM Saleh Masum

Interested in collaborating or contributing to this research?

Get in Touch →