How Canberra's Digital Archives Ended Up Full of Duplicates — and What's Being Done About It

ACT government agencies are sitting on digital image libraries riddled with duplicates — the accumulated result of more than 20 years of uncoordinated file storage, repeated system migrations, and the kind of bureaucratic sprawl that tends to flourish in a city built around public administration. The problem is not new, but pressure to fix it has sharpened considerably in 2026.

The issue matters now because of money and compliance. The ACT Government's Digital Strategy, which sets out data management obligations for agencies including Access Canberra, the Department of Health, and the Chief Minister, Treasury and Economic Development Directorate, explicitly requires agencies to eliminate redundant data holdings as part of broader cloud migration programs. Agencies that fail to do so before migrating to centralised platforms risk carrying duplicate files — and their associated storage costs — into the new infrastructure at taxpayer expense.

How the Backlog Built Up

The roots of the problem stretch back to the early 2000s. As individual directorates set up their own shared drives on servers housed variously in Civic and Fyshwick, image files — photographs from public events, scanned planning documents, maps of growth corridors in Gungahlin and Belconnen, promotional materials for programs like the now-completed Light Rail Stage 1 — were saved without consistent naming conventions or metadata tagging. When those servers were eventually consolidated, files came with them, often in multiple copies.

Staff turnover made things worse. The Australian Public Service Commission has consistently reported high rates of contractor and secondee movement across the Canberra workforce, meaning institutional knowledge about where files came from and whether a copy already existed rarely survived more than one or two budget cycles. By the time an agency realised it held three versions of the same infrastructure photograph from the Northbourne Avenue corridor, the original context for each copy had long since walked out the door.

The University of Canberra's Centre for Creative and Cultural Research documented a related phenomenon in its 2023 review of ACT cultural institution digitisation programs: without enforced deduplication protocols at the point of ingest, image collections roughly doubled in volume every four to five years even when the underlying subject matter remained static. The same dynamic applied, informally, across government.

The Push to Clean Up

Efforts to address the problem have accelerated since the ACT Government signed a whole-of-government cloud services agreement in late 2024. Under that arrangement, storage costs are metered — unlike the old on-premise model where a full server and a half-empty one cost roughly the same to run. Suddenly, duplicate images have a visible dollar figure attached to them.

The Australian National University's Research Data Management team, which advises several ACT government bodies on records practices, has been among those pushing for automated deduplication tools to be embedded in new content management system deployments. The process involves software scanning image files using hash-based comparison — essentially generating a unique fingerprint for each file and flagging matches — rather than relying on staff to manually identify copies. Several agencies piloted this approach during a trial program centred on the Majura Business Park data facility in 2025.

Results from comparable programs in other jurisdictions suggest the scale of potential gains is significant. The New South Wales Government reported in its 2024-25 Digital Government Annual Report that deduplication exercises across three pilot agencies reduced image storage volumes by between 30 and 45 percent. ACT agencies have not yet published equivalent figures, but procurement documents lodged with the ACT Procurement and Capital Works office indicate that savings targets for the current migration round are substantial.

For agencies and the public servants who work in them — many of whom live in the housing estates of Gungahlin or near the light rail corridor on Flemington Road — the practical upshot is straightforward. Staff uploading new images to agency systems will increasingly encounter automated prompts flagging potential duplicates before a file saves. Older backlogs are being scheduled for programmatic review directorate by directorate through the remainder of 2026 and into 2027. The work is unglamorous, but in a city where government is the economy, getting the digital housekeeping right has consequences that reach well beyond any individual filing cabinet.