How Canberra's government agencies ended up drowning in duplicate images — and what they're doing about it

The ACT government's digital asset libraries contain hundreds of thousands of duplicate image files. That is not an estimate plucked from thin air — it is the working conclusion that digital records managers across multiple Civic-based agencies reached after audits conducted through late 2025 and early 2026 revealed storage repositories bloated with identical or near-identical photographs, infographics, and scanned documents filed under different names, in different folders, by different teams.

The duplication problem is not new, but the scale of it — and the cost of carrying it — has finally become impossible to ignore. Two forces collided to make 2026 the year agencies could no longer shrug it off: a sharp rise in cloud storage fees as agencies migrated away from on-premise servers, and a Territory Records Act compliance review that flagged redundant holdings as a records-management liability, not just a housekeeping inconvenience.

How the problem accumulated over two decades

The roots go back to the early 2000s, when Canberra's federal and territory agencies each built separate digital asset management systems rather than shared infrastructure. A Communications and Media team at a Department operating out of Barton might photograph the same ministerial doorstop that a separate Web Services unit covered from an office on Constitution Avenue, then both teams would upload their versions to separate SharePoint libraries with no deduplication check in place.

The COVID-19 period accelerated the pile-up dramatically. When agencies scrambled to digitise paper records and run hybrid public consultations between 2020 and 2022, scanning workflows were decentralised to home offices across Gungahlin, Belconnen, and Woden. Quality control that would ordinarily catch a duplicate at the point of ingest simply did not happen under those conditions. By some internal estimates circulating inside the ACT's Shared Services ICT directorate, digitisation volumes tripled between financial years 2019–20 and 2021–22, compressing into that two-year window more new file creation than the previous decade combined.

The Australian National University's digital humanities researchers noted the same pattern in their own collections. ANU's Chifley Library undertook an internal storage review in the second half of 2025, finding that a significant proportion of image files in at least two of its archival repositories were functionally identical to files stored elsewhere on the same network under variant filenames. The University of Canberra's institutional repository team had flagged a comparable issue by mid-2025 when preparing for a system migration.

What duplicate replacement actually involves — and what it costs

Replacing duplicates is not a simple delete-and-done exercise. Best practice, as outlined in the National Archives of Australia's Digital Continuity 2025 policy framework, requires that agencies first establish a canonical master record, confirm provenance metadata is attached to that master, and only then retire secondary copies through a documented disposal process. Skipping any step creates fresh compliance risk.

The software market for automated duplicate detection has matured considerably. Tools capable of perceptual hashing — matching images that are visually identical even if their file size or format differs slightly — are now available to government procurement panels through the Digital Marketplace at licence costs ranging from a few thousand dollars annually for small agencies to well above $100,000 for whole-of-government deployments. Several ACT agencies are understood to be in evaluation phases, though no contracts have been publicly announced as of July 2026.

The practical stakes are real. Cloud storage is not free. The ACT government's migration to Microsoft Azure, which began in earnest during 2023, charges agencies for every gigabyte held. Duplicate images do not just create administrative untidiness — they generate line items on monthly invoices that agency CFOs have started scrutinising more closely as the Territory works to contain budget pressures heading into the 2026–27 fiscal year.

For agencies prepared to move, the path forward involves three steps: run an automated deduplication scan across all repositories, conduct a human review of flagged files against metadata records, and then execute a phased disposal aligned with the National Archives' retention schedules. Agencies that have already completed Phase One of similar audits — including several based in the Woden Valley government precinct — report that storage savings of between 15 and 30 percent are achievable within the first six months. The work is unglamorous. The savings are not.