Canberra's Duplicate Image Problem: What Happens Next and the Key Decisions Ahead

The ACT Government's digital asset holdings have grown to a scale where automated duplicate image detection is no longer optional — it is a budget and governance question sitting on the desks of agency heads right now. Several Territory directorates are understood to be carrying image libraries where redundant files account for a significant share of total storage, inflating annual cloud contract costs and complicating compliance with the Territory Records Act 2002.

The timing matters. The ACT's whole-of-government ICT strategy, administered through the Shared Services ICT division on Constitution Avenue, reaches a scheduled review point in the fourth quarter of 2026. That review will determine procurement priorities for the 2027-28 budget cycle. Decisions deferred past that window face at least an 18-month delay before new funding and tooling can be deployed at scale.

Why Canberra Is Particularly Exposed

Canberra's public sector workforce is unusually concentrated. The Australian Public Service employs roughly 100,000 people in the capital, and Territory agencies layer on top of that. Both tiers routinely generate visual documentation — planning photos from Gungahlin town centre development sites, infrastructure imagery from the light rail Stage 2 corridor through Civic, satellite-derived land-use captures filed by the Environment, Planning and Sustainable Development Directorate. Each project spawns multiple versions of the same images, saved by different officers across shared drives and cloud folders without consistent naming conventions.

The Australian National University's 3A Institute in Acton, which researches responsible technology deployment, has published work on data provenance and asset management in large organisations. The core problem the institute has identified in the broader literature is that duplicate files are rarely the result of malice — they accumulate through normal collaborative workflows where staff pull images from email, SharePoint, and internal portals without checking whether a canonical copy already exists in the records system.

Canberra's housing construction boom in suburbs like Belconnen and Molonglo compounds the issue. Planning and land authorities capture site photography at multiple project milestones. Without deduplication workflows, the same aerial shot of a development at Ginninderry, for instance, might exist in a dozen slightly compressed variants across as many folders by the time a project closes out.

The Decision Points Coming Into Focus

Three choices now define the near-term path. First, agencies must decide whether to run deduplication as a one-off remediation exercise or build it into ongoing ingestion workflows — the latter is more expensive upfront but avoids repeating the same cleanup in three years. Second, procurement officers have to settle on whether an AI-assisted tool or a hash-based matching system better suits government security classifications; hash matching is simpler and more auditable, which matters under Territory Records Act obligations. Third, the question of who owns the decision — Shared Services ICT, individual directorates, or a cross-agency data governance committee — remains unresolved and is the kind of turf question that historically stalls ACT digital projects.

The University of Canberra's Institute for Governance on Kirinari Street has consulted to ACT directorates on records management before. The general guidance from records management literature is that deduplication initiatives fail most often not because of technology gaps but because of unclear ownership of the remediated archive after cleanup is complete.

Cost is real. Commercial deduplication-as-a-service platforms charge on a per-terabyte or per-asset basis, and government cloud storage rates under the whole-of-government Microsoft agreements run to material annual figures once an agency's library crosses the multi-terabyte threshold. Any directorate that defers remediation past the 2027-28 budget submission deadline will carry those inflated storage costs for at least another two fiscal years.

The practical next step for any ACT agency sitting on this problem is to commission a storage audit before September 30 — the last point at which findings can credibly feed into the ICT strategy review. That audit does not require new software: most cloud platforms already surface duplication rates in their analytics dashboards. Getting that number into writing, with an agency head's signature on it, is what converts a known operational annoyance into a fundable line item.