Marjia Siddik is an AI safety researcher who has been working in AI governance, alignment, and the societal transitions advanced AI will force since 2024. She is a PhD student in Computer Science at Trinity College Dublin, funded by the Government of Ireland Postgraduate Scholarship, researching runtime detection of harmful multi-agent AI coordination.
Her past work spans international security fellowships researching US-China cooperation on artificial superintelligence, alignment research on integrity failures in deployed AI systems, and policy work shaping Ireland's first AI governance framework for its national healthcare system. She previously co-founded Safentic, a runtime monitoring system for agentic AI behavior, and was recognized as Microsoft's youngest global MVP in Trustworthy AI.
Education
Currently
Projects
Publications
A three-pillar framework for equitable AI benefit-sharing: economic redistribution, technology transfer, and capability-proliferation limits.
A phased treaty proposal combining arms-control principles with technical enforcement mechanisms to reduce ASI-driven strategic instability.
A documentation framework for medical AI datasets, funded by the Health Research Board and conducted at the ADAPT Centre.