How Tasmania's Public Image Archive Ended Up Full of Duplicates, and What It Cost to Fix It

Tasmania's state government is now partway through a remediation project to strip thousands of duplicate images from its centralised digital asset library, a problem that accumulated quietly across more than a decade of poorly coordinated uploads from agencies including the Department of Natural Resources and Environment Tasmania, Tourism Tasmania, and Parks and Wildlife Service Tasmania.

The timing matters. With the Tasmanian Liberal government under pressure to justify technology spending amid ongoing debates over the AFL stadium cost blowout at Macquarie Point and the Marinus Link undersea cable project, even routine digital infrastructure decisions are drawing tighter scrutiny from the Legislative Council. A bloated image archive is not a scandal, but it is the kind of slow-burn inefficiency that auditors tend to cite when making a broader point about asset governance.

How the Duplication Happened

The roots of the problem go back to around 2013, when individual Tasmanian government agencies began uploading photography independently to separate shared drives and content management systems before a centralised Digital Asset Management system was introduced to consolidate them. When consolidation happened, automated migration tools pulled images from multiple legacy systems, and nobody ran a thorough deduplication pass before the assets went live. The same photograph of Cradle Mountain, taken on the same morning, might exist under four different file names, uploaded by four different officers from four different agencies, each with slightly different metadata tags.

Photographers working under contract to Tourism Tasmania, whose Hobart offices are on Liverpool Street in the CBD, were often briefed separately from those contracted to Parks and Wildlife, which operates out of facilities including the head office complex in Hobart's Macquarie Street precinct. Both sets of images entered the same downstream library with no cross-check. The result, over roughly ten years, was a library where a significant share of storage was occupied by near-identical or outright duplicate files.

State Library of Tasmania staff, who manage certain archival digital collections separately from the operational government library, identified the scale of the duplication issue as early as 2021 during a review tied to the Connecting with Collections strategy. That review flagged storage costs as a growing line item. Cloud storage pricing for government contracts is not publicly itemised, but industry benchmarks suggest enterprise-grade storage for large unstructured media libraries in Australian government contexts typically runs between $8,000 and $25,000 per terabyte annually when licensing, redundancy, and access costs are factored in.

What the Remediation Looks Like

The current project uses hash-based duplicate detection, software that generates a unique fingerprint for each image file and flags identical or near-identical matches for human review. Staff at the Digital Transformation team within the Department of Premier and Cabinet are overseeing the work, with a review process designed to preserve any duplicates that carry distinct metadata of archival value, for instance, a photograph of Wineglass Bay from 2009 that documents shoreline conditions before a specific coastal management intervention.

The process is not automatic. A reviewer must still make a call on images that are near-duplicates rather than exact copies, the same scene, slightly different framing, both potentially useful to a web content officer searching for a hero image for a Parks Tasmania webpage. That human bottleneck is where the project has moved slowly. As of this month, the review is understood to be roughly halfway through the flagged batch, though no official completion date has been publicly announced by the department.

For public servants and external communications contractors who rely on the library, including regional tourism operators in areas like the Huon Valley and the East Coast tourism corridor, the practical upshot has been a library that is gradually becoming easier to search, with better-tagged primary assets replacing redundant clutter. The longer-term lesson being drawn internally is that any future agency procurement of photography should route files through a single ingestion point with deduplication checks running before assets are published, not years afterward.