The Numbers Game: What Canberra's Duplicate Image Problem Is Really Costing Government Agencies

Canberra's public sector holds one of the largest concentrations of digitised visual records in the southern hemisphere, and a growing share of that archive is redundant. Across federal agencies headquartered in Barton, Parkes and Woden, duplicate images — photographs, scanned documents, graphic assets and satellite imagery — now account for a measurable slice of government digital storage costs, according to records management specialists who work directly with Commonwealth clients.

The issue has sharpened in mid-2026 because the Australian Public Service Commission's digital uplift agenda, running since 2023, set a July 2026 milestone for agencies to complete asset audits ahead of a planned consolidation of departmental data centres. That deadline means agencies can no longer defer the question of what to do with duplicated files sitting across legacy SharePoint libraries, NAS drives and cloud buckets simultaneously.

What the Data Actually Shows

Industry benchmarks from digital asset management providers suggest that between 20 and 40 percent of files in a typical large-organisation image repository are duplicates or near-duplicates — meaning visually identical or minimally altered versions of the same source image stored under different filenames or in different folders. For a federal department running a 50-terabyte visual archive, that translates to anywhere from 10 to 20 terabytes of redundant data. At current hyperscale cloud storage rates of roughly $0.023 per gigabyte per month on AWS S3 Standard — a tier commonly used by ACT and Commonwealth agencies — a 10-terabyte duplicate load costs an agency approximately $230 a month, or just under $2,800 a year, purely in storage. That figure compounds across dozens of agencies.

The National Archives of Australia, based on Queen Victoria Terrace in Parkes, manages ingestion from multiple Commonwealth bodies and has publicly documented the challenge of deduplication in its digital preservation frameworks. The Australian Bureau of Statistics, headquartered on Lovett Tower in Belconnen, produces large volumes of chart and infographic images that cycle through multiple internal clearance workflows — a process known to generate duplicate file proliferation when version control is inconsistent.

At the ACT government level, the Chief Digital Officer's branch has been piloting an image asset review as part of the broader Digital Canberra Action Plan 2025–2028. Sources familiar with ACT Shared Services operations — which processes digital content for multiple directorates from its London Circuit offices — describe a situation where identical imagery has been uploaded to separate directorate content management systems with no automated flag to catch the redundancy.

Why Duplicates Are Harder to Kill Than They Look

Deleting a duplicate sounds simple. It rarely is. The core complication is legal: under the Archives Act 1983, Commonwealth agencies cannot destroy records, including image files, without formal disposal authorisation from the National Archives. That means a duplicate photograph of, say, a ministerial visit to the Gungahlin town centre cannot simply be binned — it must be assessed, matched against a records disposal authority, and formally sentenced before deletion proceeds.

That process has historically been manual and slow. Automated deduplication tools — software that hashes image files and flags bitwise or perceptual matches — can accelerate the identification phase dramatically. But the authorisation step still requires a trained records manager to sign off, and that workforce is stretched. The Australian Society of Archivists has noted a persistent shortage of qualified digital records practitioners in the ACT, which constrains how fast agencies can act even when the technology exists.

For agencies approaching the July 2026 audit deadline, the practical advice from records management consultants is to prioritise deduplication within active working repositories first — Teams channels, SharePoint document libraries and shared drives — before tackling cold storage archives. Running a perceptual hash tool across live systems typically surfaces 15 to 25 percent of files as candidates for review within days. The Archives Act compliance question can then be addressed systematically on that narrowed shortlist, rather than applied to an entire unfiltered archive. For Canberra's public servants already stretched by the broader digital uplift workload, that sequencing could mean the difference between meeting the milestone and carrying the bloat — and the bill — into 2027.