Zurich's Digital Archives Grapple With Duplicate Image Crisis: What Officials and Experts Are Saying

Zurich's public digital infrastructure is quietly buckling under the weight of a problem that sounds mundane but carries real financial and administrative consequences: tens of thousands of duplicate images embedded across municipal databases, heritage archives and publicly funded research repositories. The issue surfaced formally this spring when the city's Stadtarchiv Zürich, based on Alfred-Escher-Strasse, flagged that redundant files were consuming a disproportionate share of allocated server capacity, complicating record retrieval and inflating costs for taxpayers already watching the city's IT budget closely.

The timing is not accidental. Zurich has been accelerating its digitisation agenda since 2022, pushing departments to migrate paper records into centralised systems. That push — driven in part by post-pandemic administrative reforms — created conditions where the same scanned photograph or planning document was uploaded multiple times across different platforms, sometimes by different departments with no shared deduplication protocol. The result is a sprawl of identical or near-identical files that no single authority currently has a mandate to clean up.

What the Experts Are Saying

Researchers at ETH Zurich's Institute for Information Security and Dependability have been studying the broader problem of image deduplication in large institutional repositories since at least 2024. Their working position, shared at a February workshop at the ETH main building on Rämistrasse 101, is that hash-based deduplication — a technique that assigns a unique fingerprint to each file and flags matches — is reliable for identical copies but breaks down when images are slightly resized or colour-corrected before re-upload. That second category, sometimes called near-duplicate content, accounts for a significant portion of the redundancy in cultural heritage collections specifically.

At Zurich's Zentralbibliothek on Zähringerplatz, librarians have been trialling a semi-automated review workflow since early 2026, cross-referencing image metadata against acquisition records. The library holds more than 500,000 digitised items across its various collections, and staff involved in the project have described the near-duplicate problem as the harder half of the deduplication challenge — one that still requires human judgment to resolve, particularly for historical photographs where minor tonal differences may carry archival significance.

The Swiss Federal Archives in Bern have issued informal guidance recommending that cantonal bodies adopt a unified file-naming convention and a shared deduplication log before January 2027, but that guidance carries no enforcement mechanism. Zurich city officials have so far not announced a binding timetable for compliance, leaving individual institutions to manage the problem within their own operational budgets.

Costs and the Path Forward

Municipal IT storage costs in Zurich rose by roughly 14 percent between 2023 and 2025, according to the city's published ICT expenditure reports, with unstructured data growth cited as a primary driver. While duplicate images are only one component of that growth, archivists at the Stadtarchiv have estimated internally that a systematic deduplication pass across the core municipal image database alone could recover meaningful capacity — though no specific figure has been made public.

The housing shortage is an unlikely parallel: just as Zurich's Wohnungsnot reflects a structural mismatch between supply management and demand, the duplicate image crisis reflects a governance mismatch — fast digitisation without the backend protocols to keep it orderly. Critics in the archival community point out that canton-level standards exist for physical document retention but that digital asset management remains a patchwork across Zürich's 34 administrative departments.

What happens next will likely depend on whether the city's Departement der Industriellen Betriebe, which oversees shared IT infrastructure, steps in with a cross-departmental mandate before the end of the 2026 budget cycle in December. Institutions that want to get ahead of any eventual requirement are being advised by consultants working in the sector to audit their upload pipelines now, establish clear metadata ownership for each image collection, and ensure that any deduplication tool they adopt can handle near-duplicate detection, not just exact-match removal. The Zentralbibliothek project, expected to publish preliminary findings later this summer, may offer the closest thing Zurich currently has to a working model.