Canberra's Duplicate Image Problem: How the Capital Stacks Up Against Cities Tackling Visual Data Waste

Canberra's public sector has a clutter problem nobody much likes to talk about. Across federal government agencies concentrated along Constitution Avenue and in the Barton precinct, digital asset libraries have ballooned over the past decade to the point where duplicate images — redundant photographs, scanned documents and graphics stored multiple times across departmental servers — now represent a measurable drag on storage budgets and archival accuracy. The Australian National University's digital collections team flagged the issue in internal planning cycles as far back as 2023, and the National Archives of Australia, headquartered on Queen Victoria Terrace in Parkes, has been piloting automated deduplication tools since early 2025.

The timing matters because the ACT government is midway through a broader digital records modernisation push, partly driven by the Commonwealth's own data governance reforms. Storage costs are not trivial: enterprise cloud storage for government-grade data in Australia was running at roughly $0.023 per gigabyte per month as of mid-2025, and agencies holding hundreds of terabytes of unaudited image files are beginning to feel it in their annual ICT budgets. Duplicate images compound the problem because they inflate both storage costs and the time archivists spend verifying record authenticity — a particular concern when those records include infrastructure photographs for projects like Light Rail Stage 2 or planning documentation for Gungahlin's ongoing residential expansion.

What Other Cities Are Doing

The challenge is not unique to Canberra, but the approaches being adopted elsewhere vary sharply. Singapore's Government Technology Agency rolled out a whole-of-government image deduplication framework across 22 ministries in 2024, using perceptual hashing algorithms that can identify near-identical images even when file names and metadata differ. Wellington's Department of Internal Affairs has taken a more manual route, embedding records specialists in agency teams rather than automating the process — an approach that suits smaller volumes but scales poorly. Amsterdam's municipal archive, the Stadsarchief, made headlines in 2025 when it publicly reported eliminating roughly 1.4 million duplicate digital files from its holdings through a two-year AI-assisted audit, freeing up storage equivalent to around 11 terabytes.

Canberra's situation sits somewhere between Singapore's centralised automation and Wellington's hands-on model. The National Archives is trialling deduplication software across a subset of its photographic holdings — specifically digitised images from the Holt and McMahon-era collections — while the ACT government's Digital Strategy and Services division has been scoping a whole-of-territory approach that could eventually cover agencies from Health to Transport Canberra. The Australian Institute of Health and Welfare, based in Acton, is separately running its own data quality audits that include image file deduplication as one component of a broader 2025-26 data hygiene program.

What Canberra Does Differently

The capital's unusual workforce profile shapes how this problem gets addressed. With a high proportion of permanent public servants rather than contracted staff, agencies here tend to retain institutional memory about where image files live and why duplicates were created in the first place — often because of machinery-of-government changes that split or merged departments and left orphaned file directories behind. That is a genuine advantage over cities like Sydney or Melbourne where frequent contractor turnover has left large banks of unattributed digital assets with no clear custodian.

ANU's Scholarly Information Services team in the Chifley Library building has been testing open-source deduplication tools developed by the Internet Archive since February 2026, applying them to research image datasets before broader adoption. The university expects to complete its first full audit of duplicated visual assets across its institutional repository by the end of the 2026 calendar year.

For public servants and researchers dealing with this day to day, the practical advice coming out of the Archives is consistent: conduct a file-tree audit before any system migration, apply perceptual hashing rather than relying on file-name matching alone, and document the deduplication process so that audit trails remain intact. Singapore's 2024 framework — publicly available through GovTech's open documentation portal — is increasingly being cited in Canberra planning documents as a reference model worth adapting rather than simply copying.