Rethinking trust AI visibility in public services.

This article argues that trust in AI-mediated public services depends not only on whether systems are accurate, fair, or efficient, but on how their presence is made visible to the people affected by them. Across welfare, healthcare, and border governance, artificial intelligence increasingly shapes decisions about access, movement, support, and rights. Yet these systems often operate within familiar institutional procedures, making their role difficult to see, interpret, or challenge. The article develops the idea of AI visibility to explain how different forms of disclosure shape public trust. It distinguishes between invisible AI, partially visible AI, and fully disclosed AI, showing how each can produce different responses: misplaced confidence, confusion, suspicion, or informed challenge. By shifting attention from technical performance alone to the public experience of decision-making, the article shows that trust is shaped by whether people can understand where authority lies and how decisions are made. Foregrounding visibility helps reframe debates on algorithmic governance, institutional legitimacy, and accountability. In an increasingly automated state, the question is not only whether AI systems can be trusted, but whether their role in public decisions can be meaningfully seen. Working paper (Version 1.0). This manuscript has not yet undergone peer review.