Zurich's Digital Archives Are Full of the Same Photo Twice — Officials and Experts Want That Fixed

Zurich's public institutions are sitting on millions of digitised images — and a growing share of them are exact or near-exact copies of each other. That is the consensus emerging from archivists, digital preservation specialists and municipal IT managers who have spent the past several months reviewing the scale of the problem across the city's major collections. The issue is not trivial: redundant image files consume server space, inflate licensing costs, slow down search tools and, in some cases, distort the historical record when mismatched duplicates get catalogued under different metadata tags.

The timing matters for Zurich specifically. The city's Stadtarchiv on Neumarkt completed a major digitisation push in late 2025, adding roughly 400,000 new scans to its online portal. ETH Zurich's library system, headquartered on Rämistrasse, simultaneously expanded its e-rara platform for rare books and historical photographs. When two large collections grow quickly and in parallel, without a shared deduplication protocol, the overlap compounds. Conservationists and database managers at both institutions have flagged the redundancy issue in internal reviews, though neither organisation has published a formal audit to date.

What the Specialists Are Saying

Digital preservation experts at ETH Zurich's Institute for Information Systems have been studying perceptual hashing — a technique that identifies visually similar images even when file names, formats or resolutions differ — as a candidate solution for Swiss public archives. The method is already standard practice in commercial content moderation but has seen slower uptake in cultural heritage contexts, partly because archivists worry about false positives flagging genuinely distinct historical photographs as redundant.

The Zentralbibliothek Zürich, which holds one of the largest photographic collections in the German-speaking world and operates from its main building on Zähringerplatz, has been working since January 2026 with a working group convened by the Swiss Federal Archives in Bern to draft voluntary deduplication guidelines. The working group's interim recommendations — circulated to member institutions in April 2026 but not yet publicly released — reportedly favour a hybrid approach: automated flagging followed by human curatorial review before any file is removed or merged.

Municipal IT officials in the city's Departement der Finanzen have a more immediate concern: storage costs. Zurich's central government cloud contract, renegotiated in 2024, runs until 2028 and caps the city's annual data storage expenditure at a fixed rate. Duplicate image files eat directly into that allocation. While the precise storage volumes involved are not publicly disclosed, industry benchmarks suggest that unmanaged digitisation projects commonly carry duplication rates of between 15 and 30 percent of total file counts — a range that, applied to a collection of 400,000 scans, would mean tens of thousands of redundant files.

A Practical Path Forward

The Wohnungsnot housing crisis and the UBS-Credit Suisse integration have dominated the city's administrative bandwidth for much of 2025 and 2026, and digital archive housekeeping has struggled for political attention. But the direct democracy machinery that defines Swiss civic life may yet give the issue a platform. A cantonal parliamentary motion filed in March 2026 by members of the Kantonsrat called on the cantonal government to develop a unified digital asset management strategy covering all state-funded cultural institutions by the end of 2027. The motion passed its first reading and is currently in committee.

Archivists at the Stadtarchiv have said publicly — in presentations at the 2026 Swiss Archivists' Conference held in Basel in May — that any workable solution will require institutions to agree on shared metadata standards before they can meaningfully compare collections for duplicates. Without that foundation, automated deduplication tools produce too many errors to be trusted at scale.

For residents and researchers who rely on platforms like the Stadtarchiv's online portal or ETH's e-rara for historical image searches, the practical near-term advice is straightforward: cross-check results across multiple databases rather than assuming any single platform holds a clean, complete and non-duplicated record. The working group in Bern is expected to publish its full guidelines by the end of 2026, which would give institutions like the Zentralbibliothek Zürich a formal framework to work from heading into 2027.