How Canberra's Digital Archives Ended Up Flooded With Duplicate Images — And What Happens Next

Thousands of duplicate images are sitting inside the digital storage systems maintained by ACT government agencies, the result of more than a decade of overlapping digitisation drives that were never properly coordinated. The problem is not new, but it has grown urgent enough that Territory Records Office staff have been working through 2026 to develop a replacement and deduplication framework that agencies across Civic and the broader ACT public service are now expected to adopt.

The timing matters. Storage costs have climbed sharply since 2022 as cloud infrastructure pricing shifted, and the ACT government's own Digital Strategy — released in late 2023 — set an explicit target of reducing redundant data holdings across agencies by 30 per cent before the end of the 2026–27 financial year. Duplicate image files, which can represent scanned planning documents, heritage photographs, infrastructure records and social services case files, account for a disproportionate share of that bloat.

A Problem Built Over Many Years

The duplication issue has its roots in how digitisation was funded and managed across the ACT public service from roughly 2010 onward. Individual directorates — including what is now the Environment, Planning and Sustainable Development Directorate, based on Macarthur Avenue in Civic, and the older records holdings at the Noel Butlin Archives Centre at the Australian National University on Acton Peninsula — each ran their own scanning programs, often without checking whether source material had already been captured elsewhere.

The Territory Records Office, which sits within the Chief Minister, Treasury and Economic Development Directorate, was aware of the problem by at least 2018, when an internal audit flagged it as a medium-priority risk. Budget pressures shelved a proposed remediation project at that time. The issue compounded as agencies migrated to newer content management platforms — including a government-wide shift to cloud-hosted systems between 2020 and 2023 — and bulk file transfers carried duplicate images straight into the new environments without any cleansing step.

Suburban growth has added to the volume. The rapid expansion of Gungahlin and Belconnen over the past decade generated large quantities of planning and infrastructure documentation. Scanning workflows set up for those projects were often one-off arrangements, and the resulting image files were deposited into multiple repositories simultaneously as a precaution against data loss — a sensible instinct that created its own archival headache.

What the Replacement Framework Involves

The current remediation effort centres on a duplicate image replacement protocol that the Territory Records Office circulated to agency records managers in March 2026. The protocol sets out a three-stage process: automated hash-matching to identify identical files, human review of near-duplicate images where metadata differs but visual content is the same, and a controlled deletion or consolidation step that preserves at least one master copy in the authoritative record system.

Agencies have until 31 October 2026 to complete Stage One — the automated matching — across their primary holdings. The ACT Library Services branch, which manages the Canberra public library network including the Alexander Maconochie Centre library and the Belconnen and Tuggeranong branches, completed its own deduplication review in April and reported removing more than 14,000 redundant image files from its local history collection alone, according to figures the directorate published on its website in May.

For agencies still mid-process, the practical advice from records managers is straightforward: do not wait for a whole-of-government platform fix before beginning local triage. The Territory Records Office guidance explicitly allows agencies to use open-source hash-matching tools on local copies before uploading results to the central deduplication register. Teams at the University of Canberra's library and information management program on Kirinari Street in Bruce have been involved in training workshops for agency staff running that process.

The broader lesson from the ACT experience is one that other Australian jurisdictions are watching. Getting to this point took a combination of underfunded coordination, siloed procurement decisions, and a reasonable but poorly managed instinct to keep redundant backups. Fixing it requires neither dramatic technology nor large budgets — mostly patience, clear protocols, and someone being assigned to actually do the work.