Zurich's Duplicate Image Problem: The Key Decisions That Will Shape What Comes Next

Zurich's public institutions are sitting on a problem that has quietly compounded for years: tens of thousands of duplicate digital images clogging archive systems, inflating storage costs and complicating the city's push toward a unified digital infrastructure. The question now is not whether to act, but who decides how — and who pays.

The issue has sharpened this summer because Stadtarchiv Zürich, the city's official records office on Neumarkt, is entering the final planning phase of its 2024–2028 digital consolidation programme. The programme was designed partly to address redundant data accumulated when multiple city departments digitised physical records independently, often without shared metadata standards. Duplicate images — identical or near-identical scans tagged differently across systems — have emerged as one of the most stubborn outputs of that fragmented process.

Why This Moment Matters

The timing is not incidental. ETH Zurich's Image Archive unit, based on Rämistrasse, has been piloting automated deduplication tools since early 2025 using machine-learning models trained to detect near-duplicate photographs within large institutional collections. A decision on whether to scale those tools city-wide — or license them to Stadtarchiv and other municipal bodies — is expected before the end of the third quarter of 2026.

That decision carries financial weight. Storage costs for Swiss public institutions have risen alongside broader cloud pricing shifts in Europe, and maintaining redundant image libraries is no longer a trivial line item in municipal IT budgets. Zurich's 2026 city budget, approved by the Gemeinderat in December 2025, allocated CHF 4.2 million to digital infrastructure across cultural and archival departments — a figure that procurement officials say leaves limited room for unplanned deduplication projects if institutions cannot agree on a shared platform.

The housing sector has added its own layer of urgency. Zurich's Wohnungsnot crisis has pushed the city's housing office, the Amt für Wohnungswesen, to digitise inspection records, property imagery and building documentation at speed. The result, according to publicly available audit summaries from the office's 2025 annual review, is a property image database with an estimated duplication rate of around 18 percent — meaning roughly one in five images exists in at least two places in the system, often with conflicting file names and inconsistent timestamps.

The Decisions Ahead

Three questions are now in front of city officials, and none of them has a clean answer. First: should deduplication be handled centrally, through a single city-commissioned tool, or should each department manage it internally? Central coordination would be cheaper and more consistent, but it requires departments — from the Baugeschichtliches Archiv on Neumarkt to the Stadtbibliothek on Zähringerplatz — to agree on shared metadata standards they have historically resisted.

Second: what happens to images that are flagged as duplicates but differ in subtle ways — a slightly different crop, a different compression rate, a different date stamp? Archivists have flagged that algorithmic deduplication tools, including ETH's pilot model, can misidentify historically distinct images as duplicates when visual similarity is high. A manual review layer would add cost and time. Skipping it risks permanent data loss.

Third: who bears liability if images are deleted in error? Swiss data law under the revised Datenschutzgesetz, which came into full effect in September 2023, places obligations on public bodies to maintain accurate and recoverable records. Erasing a legally significant image — a building inspection photograph tied to a property dispute, for example — could expose the city to administrative challenge.

Stadtarchiv is expected to publish a formal consultation document by September 2026, inviting input from cultural institutions, legal offices and IT procurement bodies across the city. That document will not make the final call, but it will shape the options on the table when the Stadtrat votes. For institutions already stretched by the housing crisis, banking-sector restructuring and climate-infrastructure spending, the window to get this right without significant additional budget is narrow — and it is closing faster than most city officials have publicly acknowledged.