Canberra's Digital Archive Crisis: What Officials, Experts and Key Figures Are Saying About Duplicate Image Replacement

Federal and territory agencies holding millions of digitised photographs, planning maps and heritage images are under mounting pressure to address a sprawling duplicate image problem that experts say is wasting storage budgets, slowing public access portals and undermining the integrity of official records. The issue sits squarely on Canberra's doorstep: the ACT hosts a disproportionate concentration of government archives, university image libraries and planning databases, making the capital one of Australia's most exposed cities when cataloguing systems go wrong.

The push has sharpened in recent months as the National Archives of Australia, based on Queen Victoria Terrace in Parkes, has expanded its digitisation program and the ACT government's planning directorate has pushed more spatial imagery online ahead of light rail Stage 2 environmental assessments. Both efforts have surfaced the same recurring complaint from records professionals: automated ingestion pipelines routinely create near-identical copies that differ only in file name, compression level or scan date, and removing them without losing genuine originals requires specialist intervention that most agencies haven't budgeted for.

Why Canberra Agencies Are Particularly Exposed

The Australian National University's digital collections team at the Chifley Library on the Acton campus manages tens of thousands of historical images tied to research projects across more than a dozen active faculties. Staff there have been working through a backlog of suspected duplicates since at least early 2025, according to publicly available project documentation posted on the ANU library website. The University of Canberra, whose Bruce campus hosts its own special collections unit, faces a parallel challenge with images ingested during a 2022–23 digitisation grant funded through the National Library's Trove partnership program.

Records and information management professionals contacted for background this week described the core technical dilemma without ambiguity: replacing a duplicate image sounds simple until you discover that two files flagged as identical may carry different embedded metadata — geotags, rights statements or provenance chains — that make one archivally distinct from the other. Deleting the wrong version can quietly sever a link in a legal chain of custody, a particular concern for planning records that may be tendered as evidence in ACT Civil and Administrative Tribunal proceedings.

The ACT government's digital records framework, updated under the Territory Records Act 2002 and its subsequent amendments, nominally requires agencies to maintain a single authoritative version of each record. Whether operational systems actually enforce that requirement is a different matter. Territory Records Office guidance circulated in March 2026 urged directorates to conduct image audits before migrating to the new whole-of-government cloud storage environment, a rollout scheduled for completion by December 2026.

What Needs to Happen — and When

Experts in digital preservation broadly agree on a three-step approach: hash-based deduplication to identify byte-for-byte copies first, perceptual hashing for near-duplicates that differ in compression or resolution, and then human review for anything the automated passes flag as uncertain. The problem is cost. A 2024 report from the Digital Preservation Coalition estimated that remediation work on a collection of around 500,000 images — roughly the scale of a mid-sized federal agency's holdings — could take between 18 months and three years depending on metadata complexity. No equivalent Australian government figure has been publicly released.

For Gungahlin and Belconnen suburban development files, where aerial survey images are routinely re-ingested every time a planning amendment is gazetted, the duplication rate can be especially high. Planning professionals working in those growth corridors have raised the issue through forums including the Planning Institute of Australia's ACT chapter, which held a session on digital records management at its Marcus Clarke Street offices in April 2026.

The practical advice from records specialists is straightforward: agencies should not wait for a full system migration to start deduplication work, because migrating a bloated, duplicated library simply transplants the problem into a new environment at greater cost. Establishing a clear policy on which file version is authoritative — original scan resolution, rather than any derivative — and documenting that decision in a register is the minimum step any agency can take before the December 2026 cloud deadline arrives. For public servants in Canberra whose entire career output may sit inside these systems, that deadline is closer than it looks.