Duplicate Images in Canberra's Digital Records: What Happens Next and the Key Decisions Ahead

A growing pile of redundant digital images is forcing Canberra's public sector agencies and research institutions to confront a problem they have largely deferred for years: what to do with tens of thousands of duplicate files eating into storage budgets, slowing archival systems, and creating compliance headaches under the ACT's digital records management framework.

The issue has sharpened in 2026 as agencies prepare for the next round of whole-of-government ICT procurement, with the ACT Government's Digital Strategy unit expected to finalise updated data governance guidelines before the end of the third quarter. Decisions made in the coming weeks will shape how the territory manages its digital assets for at least the next five years.

Canberra is an unusual case nationally. The concentration of federal and territory agencies — from the ACT Health Directorate on Canberra Avenue to the massive digital libraries managed by the Australian National University in Acton — means the volume of image files generated locally dwarfs most comparable Australian cities. ANU's institutional repository alone holds research image datasets running into the petabyte range, and the university has been trialling automated deduplication tools since early 2025 as part of its Library Digital Preservation Program.

Why the Backlog Got This Bad

The short answer is procrastination compounded by growth. Gungahlin and Belconnen have both seen significant expansion of ACT government service delivery points since 2020, each generating local records — scanned documents, site photographs, planning images — that frequently duplicate material already held centrally on Shared Services ICT infrastructure in Macquarie. Migration projects that were supposed to consolidate those records kept getting pushed down the priority list whenever budget cycles tightened.

The University of Canberra's Faculty of Arts and Design, based on Kirinari Street in Bruce, faces a version of the same problem in its creative-practice research archive, where student project images uploaded across multiple platforms between 2018 and 2024 have never been formally reconciled. UC has acknowledged the issue internally but has not yet committed to a remediation timeline.

On the federal side, agencies clustered in Barton and Parkes — including several that shifted to hybrid working arrangements after 2021 — created secondary image repositories on departmental SharePoint tenancies that were never integrated with central National Archives of Australia guidance on deduplication. The National Archives updated its digital preservation policy in March 2024, and compliance reviews are ongoing.

The Decisions That Cannot Wait

Three choices are coming to a head before the end of 2026. First, ACT agencies must decide whether to adopt a centralised deduplication tool procured through the whole-of-government panel or allow individual directorates to source their own solutions — a fork in the road with significant cost implications either way. Centralised licensing typically reduces per-agency spend but requires governance overhead that smaller directorates often struggle to absorb.

Second, institutions like ANU and UC need to resolve ownership and retention questions before they can safely delete anything. An image that looks like a duplicate may be a different file version with embedded metadata that carries legal or research significance. Getting that determination wrong risks breaching the Archives Act 1983 or internal research data management policies tied to Australian Research Council grant conditions.

Third, and most practically, someone has to pay for the remediation work. Storage costs on the ACT government's whole-of-government cloud arrangement — managed under a contract renewed in late 2024 — are not trivial, and the bill for retaining unnecessary data falls ultimately on the territory budget. Prioritising deduplication is an argument that ICT managers have struggled to win in competition with front-line service delivery demands.

The most likely near-term outcome is a staged approach: a pilot deduplication exercise run across two or three ACT directorates in the September quarter, with results used to build a business case for broader rollout in the 2027-28 budget cycle. For ANU and UC, expect formal policy updates before the end of the 2026 academic year as both institutions push to meet updated research data management benchmarks. Agencies that defer the decision risk facing much larger remediation costs — and harder compliance questions — the longer the duplicate backlog compounds.