Canberra's Duplicate Image Problem: The Key Decisions That Will Define the Fix

A growing number of ACT government departments and Canberra-based institutions are sitting on digital asset libraries riddled with duplicate images — redundant files that inflate storage costs, slow creative workflows, and create legal exposure when licensing records can't confirm which version of an image was actually cleared for use. The problem is not new. The decisions about what to do next are overdue.

The pressure to act has sharpened in mid-2026 for several reasons. The ACT Government's Digital Strategy, which set broad data-management benchmarks for agencies to meet by June 2026, has pushed corporate services teams inside places like the ACT Health directorate on Bowes Street in Phillip and the Environment, Planning and Sustainable Development Directorate in Macarthur House to formally audit what they hold. For many, that audit is exposing the scale of the duplication for the first time.

Why This Matters More Than It Might Look

Storage is not the only issue, though it is a real one. Enterprise cloud storage contracts for mid-sized ACT agencies can run to tens of thousands of dollars annually, and duplicated image files — particularly high-resolution photography shot for infrastructure projects such as the light rail Stage 2 works along Commonwealth Avenue — can consume disproportionate capacity. Beyond cost, the risk sits in version control. When a communications team publishes an image to the Canberra.gov.au website and cannot confirm whether it is the licensed original or an uncleared duplicate, that is a liability question, not just a housekeeping one.

The Australian National University's Digital Collections team in Acton and the University of Canberra's library services in Bruce have both grappled with similar problems in their publicly accessible image repositories. Both institutions manage substantial photographic archives — UC's goes back decades and includes construction photography of the original Belconnen campus — and both have been working through deduplication projects that involve human decision-making at every contested file, not just automated matching tools.

The deduplication software question is where the sharpest decisions now sit. Automated tools can flag near-duplicate images with a high degree of accuracy, but they cannot make the call on which file to keep when two versions differ only slightly — one cropped, one with adjusted colour grading, both plausibly the correct production asset. That judgment requires a person, and that person costs money and time.

The Decisions That Cannot Be Delegated to an Algorithm

Industry guidance from the Digital Transformation Agency, which operates out of its Canberra offices and sets standards for Commonwealth entities, recommends that agencies establish a clear retention hierarchy before running any deduplication pass: original raw files sit above processed versions, licensed assets sit above internally produced ones, and files with confirmed metadata sit above those without. Agencies that skip this step and run bulk deletion risk removing the legally defensible copy and keeping the one with no provenance trail.

For Gungahlin and Belconnen community programs, where local government communications teams produce regular social content featuring suburb-specific photography, the practical next step is likely a controlled audit using a staged approach — freezing new uploads, running a hash-based duplicate check, then manually reviewing the flagged pairs before any deletion. That process, done properly for a library of several thousand images, typically takes weeks, not days.

The ACT Government's Shared Services ICT division, which manages digital infrastructure for a range of directorates from its facilities in the Canberra CBD, is positioned to coordinate a territory-wide framework for how agencies handle this. Whether it moves to do so — or leaves each directorate to solve the problem independently — is the central institutional question for the second half of 2026.

Agencies that move first on structured deduplication policies will be better placed when the next digital asset procurement cycle comes around, likely in late 2026 or early 2027. Those that don't will carry the duplication problem — and its costs — into whatever new system they buy, which tends to make a manageable housekeeping problem into an expensive migration headache.