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

The problem did not arrive overnight. Across the Australian Public Service's sprawling digital infrastructure — spread across Barton, Parkes and the suburban office towers of Woden — federal agencies have spent years quietly accumulating millions of duplicate image files, a byproduct of two decades of rushed digitisation, machinery-of-government changes and a remote-work transition that left data governance as an afterthought.

The issue matters now because the Albanese government's successor administration has prioritised a whole-of-government digital asset audit under the Australian Government Architecture framework, administered through the Digital Transformation Agency on Mort Street in the CBD. With agencies under pressure to cut storage costs and meet new data-sovereignty requirements, the duplicate-image problem has moved from an IT team's complaint to a budget line item that senior executives can no longer ignore.

How the duplication built up

Three distinct waves created the current mess. The first came between 2003 and 2012, when agencies began scanning physical archives en masse. The National Archives of Australia, based in Mitchell on the Canberra side of the ACT border, ran parallel digitisation projects with the Department of Home Affairs and the former Department of Human Services — often with no shared naming convention and no deduplication protocol. The same document could end up saved as a TIFF, a compressed JPEG and a PDF-embedded image inside the same network folder.

The second wave followed a series of machinery-of-government changes after 2013, when the Abbott government restructured 18 departments. Each merger pulled shared drives together without a consolidation plan. Content Management Systems were stitched together rather than rebuilt, and image libraries from predecessor agencies simply stacked on top of one another. The Department of Employment and Workplace Relations, for example, absorbed assets from at least three predecessor bodies whose internal photo libraries had never been rationalised.

The third and most recent wave was pandemic-driven. When the APS shifted to remote work in March 2020, staff began downloading and re-uploading images to Microsoft Teams channels, SharePoint sites and personal OneDrive folders to keep projects moving. By mid-2021, agencies were reporting that some shared communication asset libraries had tripled in apparent size — while containing little new content.

The local footprint of a national problem

Canberra feels this more acutely than any other Australian city because so much of the APS workforce is physically concentrated here. The Australian Bureau of Statistics campus in Belconnen and the Department of Finance offices in One Canberra Avenue, Forrest, both serve as examples cited in internal ICT briefings of agencies that have run their own ad-hoc image cleanup exercises — each using different tools, none sharing results with the other.

The Australian National University's 3A Institute, which has worked on responsible data systems design, and the University of Canberra's Faculty of Science and Technology have both contributed to broader thinking on digital asset governance, though neither has a formal contract role in the current audit process.

The Digital Transformation Agency published its updated Whole of Government Content Policy in February 2025, which for the first time explicitly named duplicate digital assets as a storage and security liability. Storage costs across the APS for unstructured data — the category that includes image files — were running at roughly $340 million annually as of the 2024-25 Budget, a figure drawn from the Department of Finance's ICT investment summary published alongside the Budget papers.

The remediation approach being rolled out across agencies from July 2026 involves three steps: automated hash-matching to identify byte-identical duplicates, human-in-the-loop review for near-duplicates, and a mandatory 12-month retention window before any file flagged as redundant can be deleted. Agencies in the Woden Valley precinct, including the Department of Social Services on Bowes Street, Phillip, are among the first cohort required to complete the initial automated scan by 30 September 2026.

For public servants in Canberra's growth suburbs of Gungahlin and Belconnen who routinely work from home and sync files across multiple devices, the practical implication is a likely round of prompts from agency IT teams asking staff to consolidate personal cloud storage back into official systems. It is unglamorous work, but the alternative — paying to store the same photograph of a ministerial handshake in forty-seven slightly different file formats — has stopped being a joke and started showing up in departmental audit findings.