Industry Insights / Museums, Civic & Cultural Properties
AI Visualization Requires Curatorial Boundaries
AI-assisted visualization can support interpretation when curatorial authority, cultural context, evidence, provenance, accessibility, community participation, and unmistakable labels govern every representation.
A museum may use AI to explore how something could be seen, but only curatorial evidence and community authority can establish what the institution should represent as known.
Article contents5 sections
Article navigation
Contents
The missing color returned too easily
A generated reconstruction restored color to an object whose surviving surface was fragmentary. The result helped visitors notice form and pattern. It also concealed scholarly uncertainty and the cultural authority involved in interpreting what was absent.
The image looked complete because the evidence was not.
Curation governed the purpose
ICOM's museum definition places research, interpretation, accessibility, inclusion, ethics, and community participation inside the institution's public role. NIST's AI framework adds tools for considering validity, transparency, and harm.
Together they kept AI use subordinate to a defined interpretive purpose and a qualified human decision.
Provenance made uncertainty visible
Source objects, scholarship, community knowledge, prompt inputs, model use, material changes, rights, and approval all belonged in the record. Labels needed to explain what was reconstructed and what remained uncertain in language and formats visitors could access.
Some representations should not be generated at all when cultural restriction, dignity, contested meaning, or likely harm outweighed interpretive value.
Evaluation included visitor understanding, not only visual appeal. If audiences consistently read a hypothesis as fact, the label or presentation had failed even when technically present. Curatorial teams needed authority to revise or remove a representation after publication.
The concept stayed inside interpretation
SiteSee can connect an AI-assisted visualization to location, evidence, labels, versions, access information, and curatorial approval. It extends the spatial-storytelling standard without converting uncertainty into scenery.
The system preserves the boundary. Curators, scholars, and communities determine where it belongs.
The fragment kept its authority
The restored color could remain as one documented interpretation. Beside it, the surviving fragment and the limits of knowledge stayed visible.
The incomplete object no longer looked like a defect the model had fixed. It remained the evidence to which every visualization had to answer.
Sources and evidence
- International Council of Museums, “ICOM Approves a New Museum Definition” (opens in a new tab), August 2022.
- United States National Institute of Standards and Technology, Artificial Intelligence Risk Management Framework 1.0 (opens in a new tab), January 2023.
- United States National Institute of Standards and Technology, AI Risk Management Framework Playbook (opens in a new tab), March 2023.
- World Wide Web Consortium, Web Content Accessibility Guidelines 2.1 (opens in a new tab), June 2018.
- Tussyadiah and colleagues, “Virtual reality, presence, and attitude change” (opens in a new tab), Tourism Management, 2018.
Last updated: 2026 08 23