Zurich's Digital Archive Push Stalls Over Duplicate Image Problem: What Officials, Experts and Key Figures Are Saying

Thousands of duplicate images are clogging Zurich's municipal digital archives, slowing public access to historical records and driving up storage costs, according to archivists and technology specialists who have been pressing the issue with city administrators this spring. The problem sits at the intersection of two of Zurich's long-running institutional priorities: digital transformation of public services and responsible stewardship of civic memory.

The issue is not unique to one department. It spans at least three major city-linked repositories — the Stadtarchiv Zürich on Alfred-Escher-Strasse, the Zentralbibliothek on Zähringerplatz, and the photography collection maintained under the umbrella of the Präsidialdepartement. Each institution digitised holdings independently over the past decade, often without a shared metadata standard, producing overlapping catalogues where a single historical photograph of, say, the Limmatquai or the Grossmünster may appear under four or five different file entries with slightly varied descriptions.

Why the Problem Is Getting Harder to Ignore

Storage is no longer the only pressure. Under Switzerland's revised Archivgesetz framework, institutions are required to ensure retrieval accuracy and data integrity when public records are requested — a standard that becomes difficult to meet when automated search tools surface duplicate results. Specialists in digital preservation have been pointing out for at least two years that the duplication rate in large Swiss municipal image collections typically runs between 15 and 30 percent of total holdings, depending on how duplicates are defined. At those rates, a collection of 200,000 digitised items could carry 30,000 to 60,000 redundant files.

ETH Zurich's computer science and information science departments have been developing AI-assisted deduplication tools with potential public-sector applications, and discussions between ETH researchers and city archivists have been ongoing since at least early 2025. The core technical challenge is that true duplicates — identical files saved twice — are easier to catch than near-duplicates: slightly different scans of the same original photograph, or images cropped differently for separate publications. The latter category requires more sophisticated image-recognition models and is the source of most of the cataloguing headaches.

City councillors on the Gemeinderatskommission für Kultur und Bildung raised the matter in committee sessions earlier this year, with the question of budget allocation for a unified deduplication project yet to be resolved. A coordinated tender for a cross-institutional digital asset management platform was anticipated by mid-2026, though no formal award announcement had been made as of this week.

What Needs to Happen Before the Problem Gets Worse

Practitioners working in digital heritage management argue that the longer institutions wait, the more expensive a fix becomes. Each year of continued independent digitisation — without a shared identifier system — adds fresh layers of redundancy that retroactive cleaning tools will have to process. The Stadtarchiv alone added several thousand new digital items in 2025 under ongoing municipal records-management mandates.

The Zentralbibliothek, which holds one of the most extensive historical photograph collections of the city and the broader canton, has been piloting a metadata harmonisation project in its maps and prints section since autumn 2024. That pilot is seen by archivists as a potential template, though scaling it across institutions requires both political will and a central coordination body — something that does not currently exist in Zurich's decentralised administrative structure.

Practically, for researchers and members of the public who use these archives — whether at the reading room on Alfred-Escher-Strasse or through the online portal — the near-term advice from information specialists is consistent: cross-reference any image result against at least two institutional catalogues before assuming uniqueness or completeness. Until a shared platform is in place, the search burden effectively falls on the user. That is not a sustainable position for a city that regularly cites digital accessibility as a core value of its cultural policy.