DR THOMAS GRAHAM

Thomas Graham

If you would like to request a contact for this project, please email adms@rmit.edu.au

Dr Thomas Graham completed his PhD in July 2025 at Swinburne University.

Thesis Title
Responsible AI in Space: Regulating AI in the Final Frontier

Summary

Deployment of artificial intelligence (AI) in the space sector — including autonomous satellite operations, collision avoidance systems, and Earth observation analysis — presents complex challenges that the existing international space law framework cannot clearly address. This research investigates the regulatory gap in AI governance for the space sector. Through comprehensive analysis of international space law instruments, emerging AI regulatory frameworks (particularly those of the European Union and Australia), ISO standards, and Responsible AI frameworks, combined with insights from a survey of 103 space and AI sector professionals from around the world and 15 semi-structured expert interviews (primarily with Australian respondents), it seeks to locate the most promising pathways for regulating AI and promoting Responsible AI practices in the space sector.
International space law, drafted decades before modern AI, contains significant ambiguities in the area of fault attribution for opaque AI systems, thus failing to adequately address liability for damages from AI-related faults. Analysis of emerging AI regulatory frameworks suggests they offer the most effective model: binding guidance for the growing space sector.
Survey respondents and interviewees demonstrated awareness of Responsible AI concepts, but implementation remained limited; resource constraints meant they were likely to prioritise AI governance only if it were mandated. While respondents preferred voluntary standards over binding regulation, and were concerned about regulatory burden and outdated or overly prescriptive requirements, they also acknowledged the need for binding regulation in the long term.
The research suggests that binding, risk-based AI regulation, nationally and transnationally, complemented by widespread adoption of international standards and tailored education initiatives, are the best way forward. Preparing the space sector to adopt Responsible AI practices willingly, as a positive AI risk mitigation strategy rather than an imposed regulatory burden, preserves its ability to innovate, advance, and remain competitive. This research contributes to both space law and AI governance scholarship by examining a sector previously overlooked in AI governance literature, offering practical guidance for policymakers, space agencies, and private sector organisations navigating this evolving regulatory landscape.

Supervisors
Prof Anthony McCosker, Swinburne University