How Zurich's Digital Archive Crisis Reached Breaking Point: The Story Behind the Duplicate Image Problem

Zurich's municipal digital infrastructure is carrying tens of thousands of redundant image files across at least a dozen city departments, a problem administrators have been quietly working to resolve since a 2023 internal audit first quantified the scale of the issue. The findings from that review — conducted by the city's Amt für Informatik, the IT office based near Rennweg in the first district — identified duplicate imagery as a significant driver of rising storage costs and a persistent source of confusion in public-facing digital communications.

The problem did not appear overnight. It is the product of roughly fifteen years of decentralised content management, during which individual departments at institutions ranging from the Stadtarchiv on Neumarkt to the communications teams at Stadtspital Zürich built their own image libraries with no shared taxonomy, no unified metadata standard, and no cross-departmental deduplication protocol. When Zurich's central web platform migrated to a new content management system in stages between 2019 and 2021, thousands of legacy assets were imported as-is, duplicates included.

A City-Wide Problem With a Paper Trail

The 2023 audit found that roughly 34 percent of image assets stored on city-managed servers were exact or near-exact duplicates, according to figures published by the Stadtrat in a planning document circulated to the Gemeinderat that autumn. Storage costs for those redundant files were running at an estimated CHF 180,000 annually when server lease, bandwidth, and maintenance labour were factored together. That figure does not include the hidden cost of staff time: communications officers across departments reported spending an average of 40 minutes per week searching for the correct, authoritative version of a frequently used image — a photograph of the Grossmünster, say, or an aerial of the Hardbrücke neighbourhood — only to find three or four versions with conflicting file names and unclear licensing metadata.

The licensing dimension matters particularly in Switzerland. Under Swiss copyright law, which was updated via the revised Urheberrechtsgesetz that came into force in April 2020, organisations using images without clear provenance documentation face sharper liability exposure than under the previous framework. Several cantonal bodies received formal notices from photographers' rights groups in 2022 and 2023 after duplicate or improperly attributed images surfaced in official publications. Zurich was not publicly named in those proceedings, but the timing accelerated internal pressure for a systematic fix.

What Zurich Is Now Doing About It

The city's response has unfolded in phases. In early 2024, the Amt für Informatik piloted a deduplication tool across the Stadtentwicklung department's image repository, a collection of roughly 22,000 files built up since 2008. That pilot reduced the active library by 28 percent within three months and is now being used as a template for a city-wide rollout scheduled to run through the end of 2026. The ETH Zürich spin-off involved in providing the underlying image-matching algorithm — a computer-vision approach that catches near-duplicates even when file names and metadata have been altered — signed a service contract with the city in March 2025.

The Stadtbibliothek Zürich, which operates the central library on Zähringerplatz and manages a separate but parallel digital asset system for its own archival collections, is running a concurrent but independent deduplication project. The two institutions have agreed to align metadata standards by the fourth quarter of 2026, which would lay the groundwork for a shared city asset registry — something administrators have discussed since at least 2017 without previously securing the budget to pursue it.

For departments not yet included in the rollout, the practical guidance from the Amt für Informatik is straightforward: freeze new uploads to legacy repositories, document licensing for every image currently in active use, and flag duplicates manually using a checklist the office distributed in February 2026. The full city-wide deduplication sweep, once complete, is projected to recover approximately 4.2 terabytes of server capacity — modest by enterprise standards, but meaningful for a public administration working within the constraints of Zurich's annual IT infrastructure budget of CHF 47 million. The longer-term prize is a single, searchable, legally clean image commons that every city department can draw from — and that does not need to be rebuilt from scratch every time a CMS platform changes.