Syme A1,2, Price G2,3, Huynh M3,4, Goudey B2, Hyde C5, Bromhead C1,2, Thang M3, Goonasekera N1,2
1University Of Melbourne, Parkville, Australia, 2Australian BioCommons, Melbourne, Australia, 3University of Queensland, St Lucia, Australia, 4Sydney Informatics Hub, Newtown, Australia, 5QCIF, Brisbane, Australia
Biography:
Dr Anna Syme is a senior bioinformatician on the science team at Galaxy Australia, part of Australian BioCommons. She works across bioinformatics workflows and other science and platform activities, and is product owner for Genome Lab. Her current interest is how AI tools and agents can be adopted safely and usefully on national research platforms.
Abstract:
Artificial Intelligence (AI) has arrived in research computing from every direction. Researchers use chatbots and AI coding assistants, and the platforms they depend on are now building AI into the infrastructure. Galaxy, the open-source platform for accessible, reproducible data analysis and run nationally as Galaxy Australia, is a case in point. Galaxy already hosts tools that apply AI within the analysis itself, such as AlphaFold to predict protein structures. But AI is being increasingly built into the platform. Various features are under development: an in-browser conversational assistant; a Model Context Protocol (MCP) server that exposes Galaxy's tools, workflows and histories to external AI agents; and an AI agent framework. Researchers, and AI agents acting for them, now reach Galaxy through the web interface, an agent using a user's API key, or the command line. To map this, we built a diagram tracing where each feature sends a user’s data: what leaves the platform, and which model receives it. It clarifies the privacy questions that users face and serves as a governance tool, placing each feature against Galaxy Australia's Acceptable Use Policy and Service Level Agreement and flagging unresolved data handling, so AI can be adopted securely and usefully for research.
