Addressing User Safety in Software Systems Using Generative AI

With advances in technology, the range of users has expanded to diverse groups of varying ages, genders, cultures, characteristics, preferences, and working environments. The safety of software systems is a critical consideration in software design, particularly in safety-critical domains such as autonomous vehicles and aged care, where failures or poor user experiences can have serious or even life-threatening consequences.;Research Gap;Existing research has primarily focused on technical aspects and often overlooks the safety issues users face when interacting with software-only and software-intensive systems. A deep understanding of these user-facing issues can help software developers and designers create safer and more inclusive systems. Software repositories, such as GitHub and software reviews, contain valuable insights into the challenges users encounter and how developers address them. However, current works do not systematically investigate and mitigate safety requirements. ;Aim;This research project aims to support software developers and designers by identifying and addressing software safety concerns from a user-centred perspective, using guided Generative AI (such as through AI agents) on software reviews and developer discussions. ;Objectives;We expect to develop a novel Generative AI and machine learning tool to automatically detect, classify, and summarise safety issues from these data sources. The AI tool will be trained on curated datasets and fine-tuned to identify patterns related to user safety concerns, enabling scalable and proactive identification of risks in software systems. Furthermore, the research will deliver effective ways to generate architectural and design interventions that can be integrated into the original system(s) for improved safety.;This PhD project presents a unique opportunity to conduct impactful, interdisciplinary research at the intersection of human-centred software engineering and advanced AI. Jointly offered by IITM Academy and Deakin University, the project gives candidates access to world-class expertise, resources, and international collaboration. The candidate will develop and apply state-of-the-art Generative AI and machine learning techniques to analyse large-scale, real-world data. The focus will be on identifying and addressing safety-related user concerns, contributing to the development of more inclusive and responsible software systems. This work is ideally suited for candidates interested in AI-driven requirements engineering, safety assurance, and the practical application of GenAI in software development pipelines. The project offers scope for original contributions to both academic research and industry practice in a rapidly evolving and socially relevant field.;;Keywords: Safety, software, AI, machine learning, online repositories, user-centred design.;Necessary Skills:;• Python;• Knowledge of machine learning;• Familiarity with repository mining (preferred but not required);• Interest in Generative AI applications in software engineering