Charlton G1, Le L1, Liu Y1, Ashton A2, Gorringe M2
1Advanced Analytics and AI Platform, Intersect Australia, , Australia, 2Sax Institute, , Australia
Biography:
Glen Charlton is the Lead Data Scientist for the Advanced Analytics and AI Platform and oversees the Data Science services provided by Intersect Australia . He has over 10 years of experience in applying Data Science and Electronics Engineering to commercial and research projects primarily within health and environmental sciences. Glen established a research track record demonstrating the application of data science across multiple sciences with the key interest of improving the efficiency and increasing the capability of research projects to conduct novel and practically relevant research.
Matthew Gorringe leads the Data and Research Services Division at the Sax Institute, overseeing the Secure Unified Research Environment (SURE), the 45 and Up Study, and the 18 and Up Study. With more than 20 years’ experience across health survey design and implementation, large‑scale data collection, data governance, and secure research infrastructure, he brings deep expertise in managing complex data environments and research operations. Matt drives strategic initiatives that enhance data collection, participant engagement and SURE system capacity, performance, and researcher experience, delivering measurable improvements in secure linked data access for researchers. He has extensive expertise in survey research methods, data management and analysis, and the leadership of large, multidisciplinary research and operational teams. Matt also brings senior executive experience from the human services sector and holds a Graduate Certificate in Survey and Market Research Methods and a Master of International Relations.
Abstract:
Background: The Secure Unified Research Environment (SURE), managed by the Sax Institute and funded by the Population Health Research Network (PHRN) and NCRIS, is a critical national asset facilitating secure health data research. As the volume and complexity of input and output vetting grow, this poses a significant resourcing challenge to have highly qualified staff completing the task manually. Using a replica environment, we have developed a Proof-of-Concept AI-Enabled Curation Assistant to automate and streamline data curation processes within SURE.
Method: AI-Enabled Curation Assistant is a multi-stage curation pipeline that moves beyond traditional manual review and was developed as part of the SURE 3.0 Enhancement Project in response to stakeholder feedback. The system operates without external communication and utilises a hybrid execution model, combining rapid deterministic rule-based checking with specialised Large Language Models (LLMs). The pipeline systematically processes submissions through four stages: file identification, data type extraction, hybrid analysis, and a final holistic reasoning phase.
Results: AI-Enabled Curation Assistant gives researchers immediate feedback and curators a tool to assist with decision making. The Proof-of-Concept can successfully ingest and process various file and data types. This will allow for real-time feedback, reduce operational pressures and increase the data security for data custodians.
Conclusion: This approach addresses the challenges of scaling secure research infrastructure, offering a robust, scalable framework that maintains data privacy and improves operational efficiency. We will share preliminary performance benchmarking results and discuss the next phase of testing, data custodian acceptability consultation and deployment within the SURE Curated Gateway.
