AI in the Curator’s Loop: Balancing Automation & Authority

This slide deck explores an emerging design, regulatory, and trust challenge for museums: how to communicate the presence of AI-influenced content to users in line with the forthcoming EU AI Act, without undermining the long-established perception of museums as authoritative and trustworthy sources of knowledge.The presentation does not introduce an AI system or propose AI-assisted metadata as a deployed solution. Instead, it uses illustrative mock-ups and screenshots containing AI-generated example content to visualise different levels of AI notification and disclosure. These examples were derived from Project SPOT outputs solely to provide realistic, domain-appropriate material for design discussion, not to evaluate the AI itself.The focus of the work is on how AI disclosure is shown, rather than how AI generates content. It examines where, when, and how notifications of AI involvement might appear across museum interfaces, from site-wide indicators to artefact-level and content-level signals. The central concern is how such notifications may alter user trust, given that museum audiences are accustomed to assuming that published information is carefully researched, curated, and as accurate as possible.By framing transparency as a user experience and epistemic trust problem, the presentation raises questions about whether mandatory AI disclosure risks unintentionally signalling lower reliability, even when all content has been curator-reviewed and approved. It positions transparency not as a binary label, but as a design space requiring careful consideration of language, visibility, and cognitive load.The deck is intended as a discussion starter for researchers, designers, curators, and policymakers, highlighting the need for empirical research into how AI notifications affect user trust and how museums can remain compliant while retaining curatorial voice and institutional authority.