'It felt like our story had been erased': Perth residents speak out on duplicate image problem hitting community archives

Residents across Perth are raising alarms about a quiet but damaging problem eating through digitised community archives: duplicate images that, when flagged for automated removal, are taking genuine, non-duplicated photographs with them — erasing local history in the process.

The issue has surfaced repeatedly over the past six months in conversations among community groups in Fremantle, Midland, and the inner-northern suburbs. Local historical societies say they first noticed the problem in late 2025, when batch-processing software used to clean up scanned photo collections began misidentifying similar — but distinct — images as exact duplicates, triggering deletions that could not always be reversed.

Communities bearing the cost

In Fremantle, members of a long-running neighbourhood history project based out of the Fremantle Arts Centre on Finnerty Street say they lost a run of photographs documenting the 1987 demolition of a terrace row on High Street after a deduplication process flagged the near-identical sequential shots as redundant. The images, taken within seconds of each other from the same vantage point, were distinct records of a continuous event. All but one frame survived.

Similar frustrations have surfaced in Midland, where volunteers working with the Swan Valley Historical Society have been manually reviewing hundreds of scanned glass-plate negatives after discovering that automated tools applied to their collection removed what appeared to be a set of early 20th-century orchard photographs. Several community members described spending weekends cross-checking physical negatives against digital catalogues to identify what was lost and what might be recoverable from backup drives.

One long-time Bassendean resident, who has spent more than a decade contributing to local digitisation efforts along the upper Swan River corridor, described the experience as demoralising. The work of community archivists — most of them unpaid — rests on the assumption that once an image is digitised, it is safe. That assumption, many now say, was wrong.

The problem is not unique to Perth, but the city's rapid population growth — Western Australia's population crossed 2.9 million in 2025 according to the Australian Bureau of Statistics — has increased demand for digitised community records as new residents seek to understand the places they have moved into. That pressure is pushing institutions and volunteer groups to process collections faster, and faster processing means more reliance on automated tools that are not always calibrated for nuanced archival material.

What needs to happen now

The State Records Office of Western Australia, based on Alexander Drive in Osborne Park, maintains guidance on digital preservation standards, and archivists working in the sector say community groups should treat that guidance as a baseline rather than an aspiration. Specifically, experts in digital preservation recommend that any deduplication process be preceded by a full verified backup and that automated tools never be given delete permissions on primary collections without manual review of flagged items.

The Battye Library, part of the State Library of Western Australia on Francis Street in the Perth CBD, has run community digitisation programs that include basic training in file management. Participants in those programs say the training does not currently include a dedicated module on deduplication risks — a gap that community advocates are now pushing to fill.

Volunteers and local historical societies say the most practical immediate step is a moratorium on automated deletion in any community-run collection until a manual audit can confirm what the software has flagged. For groups without the technical capacity to run that audit themselves, the State Records Office offers a consultation service that community groups can access at no cost.

The broader ask is for the WA government's culture and heritage portfolio to fund a dedicated digital preservation officer role embedded with community organisations — something advocates say has been discussed informally but has not yet reached a budget commitment. With the state currently running a significant surplus off the back of iron ore royalties, the funding case is there. The political will is still being tested.