Thesis Title
Designing for responsible urban innovation with AI technologies for local government

Research Description
Local governments globally, including in Australia, are undergoing digital transformation, with automation at the forefront. AI technologies, like automated decision-making systems, generative AI, language-processing and Geo AI are being employed to enhance efficiency, accelerate urban planning, for service-delivery and to address complex challenges in the built environment . However, concerns about the responsible design, use of these technologies, community engagement and impact in local government for urban design, the built environment and placemaking are often overlooked.

A critical gap exists between responsible AI research and its real-life design and application, with valuable insights from research not reaching planners, designers, and policymakers and hindering evidence-based implementation of technologies in placemaking and local governance, while local governments’ approaches to these technologies and their social and environmental considerations remain unclear and fragmented.

‘Responsible urban innovation’ as a response to these issues foregrounds a value-sensitive approach that usefully prioritises social responsibility, environmental sustainability, and long-term solutions. The responsible urban innovation approach can integrate design-thinking and community engagement methodologies—such as co-design, co-creation, and co-evaluation—as essential processes for embedding community participation and inclusion into the development and implementation of technology-based innovations.

This project examines the practical implications of digital transformation in local government for communities and local government practitioners, by focusing on design-thinking and adoption of AI technologies in areas such as planning and design, housing, community building, and disaster preparedness. It aims to evaluate responsible and innovative practices in technology design and adoption, assess the social value for communities, and explore how design-thinking approaches enhance responsible urban innovation and build trust within organisations and communities through meaningful participation, transparency, and communication.

Supervisors
Prof Anthony McCosker, Swinburne University

Tracy is also ADM+S node adminstrator for Swinburne University. Read her node administrator profile here.

Thesis Title
Fairness in recommendation systems

Research Description
Jessica will be researching fairness in recommendation systems, particularly looking at multiple definitions of fairness.

Supervisors
Assoc Prof Jeffrey Chan, RMIT

Thesis Title
Freedom of Expression and Cyber Bullying in Democracies; Legal Responses

Research Description
Charhurika’s research explores how approaches to regulate online speech have supported people’s participation in selected commonwealth countries.

Supervisors
Prof Andrew Kenyon, University of Melbourne

Thesis Title
User-centered Non-factoid Answer Retrieval

Research Description
Marwah’s research aims to examine the assumptions that are made about users when searching for non-factoid answers using search engines. That is, the way they approach non-factoid question-answering tasks, the language they use to express their questions, the variability in their queries and their behaviour towards the provided answers. The investigation will also examine the extent to which these neglected factors affect retrieval performance and potentially guide the development of more realistic methodologies and test collections that capture the real nature of this task.

Marwah is interested in understanding how users interact with search engines and how we can use that understanding for designing user-centered information retrieval systems. She is currently investigating ways to help users answer complex information needs beyond the use of Search Engine Result Pages (SERPs). Her Ph.D. research particularly focuses on how users vary in the way they seek information for complex information needs and the impact of that variation on retrieval effectiveness. This should enable equally effective retrieval systems regardless of the observed variations among users.

Supervisors
Prof Mark Sanderson, RMIT University
Professor Falk Scholer, RMIT University
Paul Thomas, Microsoft Research