AI-slands Towards an Island-Centred Approach to Artificial Intelligence

Artificial intelligence is frequently discussed as though it were a placeless technology whose effects will spread relatively evenly through increasingly connected societies. An Island Studies perspective calls this assumption into question. AI is produced through concentrated infrastructures of capital, computing power, energy, data and specialist expertise, while its applications enter communities characterised by very different geographical, demographic, cultural and political circumstances. This paper proposes AI-slands as an island-centred framework for examining a central tension: artificial intelligence may compress some consequences of geographical distance while simultaneously increasing dependence on distant infrastructures, platforms and representational systems. The paper organises its contribution as an analytical sequence. AI archipelagos provides the spatial frame; algorithmic mainlanding names a mechanism by which metropolitan or mainland conditions become algorithmic defaults; digital double insularity and synthetic islandness describe resulting forms of representational distance and machine-produced island imagery; and insular AI sovereignty identifies a practical governance response centred on island agency rather than technological self-sufficiency. The argument is conceptual and propositional. It is intended to generate hypotheses for comparative island research, particularly through a proposed Island Representation Audit and studies of AI in island services, culture, language, governance and infrastructure. Drafting and site development for this working paper made use of AI tools; analytical claims, framework design and editorial judgement remain the author's, and sources have been checked prior to publication.