Artificial intelligence and machine learning for Australian marine science: current state and future needs

Le résumé fourni par la source

Australia’s marine estate, one of the world’s largest, supports a blue economy worth more than AU$100 billion annually across fisheries, aquaculture, tourism, offshore energy and maritime operations. Yet Australia’s capacity to monitor and manage these environments is constrained by a widening gap between rapid data collection and analytical capability. The IMOS Understanding Marine Imagery Facility alone holds about 10 million images, fewer than 3% of which have been annotated, representing up to AU$150 million in unmet manual analysis costs. Artificial Intelligence (AI) and Machine Learning (ML) can help close this gap by automating data processing, enabling real-time analysis and supporting predictive, evidence-based decision-making. This paper presents a national assessment of AI and ML in Australian marine science, based on a survey of 42 projects across 18 institutions. Results reveal a growing portfolio of applications across biodiversity conservation, ocean monitoring, fisheries management and climate science, with projects spanning Technology Readiness Levels 0 to 9 and an average target TRL of about 8. Key barriers include workforce shortages, infrastructure constraints, fragmented data systems and funding uncertainty. Addressing these challenges requires coordinated investment in people, infrastructure and data ecosystems, stronger partnerships, responsible AI governance and a long-term national research agenda for sustainable marine management.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.