Ethical Dimensions of Human-AI Interaction in Maritime Decision-Making: An Expert Perspective

Abstract

As artificial intelligence systems progressively enter maritime decision-making environments — from AI-assisted collision avoidance recommendations on bridge displays to algorithmic cargo scheduling in port operations and predictive maintenance alerts in engine rooms — the ethical dimensions of human-AI interaction in these high-consequence professional contexts demand empirical investigation that current maritime ethics scholarship has not yet provided. This study examines the ethical dimensions of human-AI interaction in maritime decision-making through qualitative focus group discussions with 30 experts across Port and Shipping Management (PSM), Deck Nautical (DN), and Marine Engineering (ENG) program affiliations at STIP Jakarta, drawing on the expert practitioners' lived experience of automation transitions in maritime contexts. Four thematic categories were generated through systematic analysis: Conditional Trust Architecture, Accountability Asymmetry, Human Override Norms, and Generational Ethics Gradient. DN experts exhibit the strongest human-primacy ethical orientation (most pronounced endorsement of override norms and conditional trust), while PSM experts exhibit the most pragmatically accepting orientation toward AI decision support. The Generational Ethics Gradient theme documents a practitioner consensus that digital-native generations entering maritime professions will hold weaker human-primacy values than their predecessors, with expert participants identifying this generational transition as the most consequential and least institutionally prepared ethical dimension of maritime AI integration. The study contributes the first practitioner-grounded, ethics-framework-informed account of human-AI interaction values in Indonesian maritime professional contexts