Zurich's Digital Archives Are Full of Duplicate Images — Officials and Experts Want Answers

Zurich's public institutions are sitting on vast repositories of duplicated digital images — and the bill is growing. City archivists, computer scientists at ETH Zurich, and data governance specialists have spent much of 2026 wrestling with what one internal assessment described as a systemic failure in how municipal and academic bodies manage photographic and visual records. The problem is not new, but pressure to act has intensified this year as the city's digitalisation push collides with budget constraints in the Stadthaus and across Zurich's publicly funded research sector.

The core issue is straightforward: when organisations digitise documents, run large-scale photography projects, or migrate legacy systems, duplicate images accumulate quietly. They consume server space, complicate legal compliance under Switzerland's revised Data Protection Act — which came fully into force in September 2023 — and erode the reliability of public records. For a city that has staked considerable political capital on smart governance and digital transparency, the redundancy problem is an embarrassment as much as a cost.

What the Experts Are Saying

Researchers at the ETH Zurich Computer Science department in the Hönggerberg campus have been developing hash-based deduplication tools capable of processing large visual datasets. The approach involves generating unique digital fingerprints for each image file; if two fingerprints match, one copy is flagged for review and potential deletion. Specialists in the field point out that perceptual hashing — which catches near-identical images, not just exact copies — is increasingly the standard demanded by institutions handling archival material at scale.

The Stadtarchiv Zürich, located near Neumarkt in the old town, manages tens of thousands of digitised records accumulated over decades of scanning projects. Archival professionals there have publicly acknowledged, in presentations at the Zurich Information Management Forum earlier this year, that deduplication protocols remain inconsistent across departments. The challenge, experts note, is not purely technical: decisions about which copy to retain, which metadata to preserve, and who holds authority to delete a record touch on questions of institutional memory and democratic accountability that Swiss direct democracy makes especially sensitive.

Kantonsrat members have raised related concerns in committee sessions about the cost efficiency of the city's data infrastructure. Server storage for the municipal administration runs into the millions of francs annually, and independent technology consultants briefed by the city in early 2026 reportedly estimated that duplicate and redundant files — images chief among them — account for a meaningful share of that overhead, though no official figure has been published.

Practical Steps and What Comes Next

The University of Zurich's Information Science faculty at Rämistrasse 74 has also entered the conversation. Faculty there have been collaborating with the Swiss Federal Archives in Bern on a framework for responsible automated deduplication — one that would give institutions a replicable workflow rather than leaving each body to improvise. A pilot program involving several cantonal departments is expected to begin trials in the third quarter of 2026, with a formal evaluation scheduled for early 2027.

For Zurich's private sector, especially the pharmaceutical and financial industries concentrated along Talstrasse and in the Zürich West innovation quarter, the stakes are different but no less real. After the UBS absorption of Credit Suisse, compliance teams at several Zurich-based financial institutions have had to reconcile enormous document archives, and duplicated image files within compliance records have created headaches during regulatory audits conducted under FINMA oversight.

The practical advice from technology specialists is consistent: institutions should audit their current storage environments before deploying any automated tool, establish clear retention policies tied to legal obligations under Swiss law, and designate a named data steward responsible for sign-off on deletions. Automated solutions can flag duplicates faster than any human team, but the final call on what constitutes the authoritative copy still requires human judgment — particularly in archival and compliance contexts.

Zurich will not resolve this in a single budget cycle. But with the city's 2027 digitalisation strategy already in draft at the Stadthaus, and ETH Zurich's deduplication research moving toward publication, the window for coordinated action is open. The question officials are now asking is whether the political will to standardise across institutions — cantonal, municipal, and academic — can be assembled before another year of redundant storage costs accumulates.