Paris Tackles the Duplicate Image Problem Reshaping Urban Digital Archives — and Other Cities Are Watching

Paris city hall quietly launched a digital archive overhaul in January 2026, targeting a problem that has ballooned across every major European municipality: duplicate images clogging planning databases, heritage registries, and public-facing cultural portals. The scale is significant. According to internal figures cited by the Direction des Affaires Culturelles de Paris — the municipal body overseeing the city's cultural assets — the shared image repository used by planning offices and heritage services had accumulated more than 340,000 redundant files by late 2025, slowing access times and muddying search results for urban planners working on Grand Paris Express integration projects along the outer arrondissements.

The problem is not trivial bureaucracy. City planners and conservationists rely on accurate photographic records to make decisions worth millions of euros. When the same storefront on the Rue de Rivoli appears under six different file names with conflicting metadata, or when a heritage photograph of a Haussmann-era façade in the 9th arrondissement is duplicated across three separate departmental databases, the downstream errors compound fast. Licensing disputes, redundant print orders, and misidentified structures have all been traced back to duplicate image records in recent audits, according to background documentation circulated within the Direction de l'Urbanisme.

What Paris Is Actually Doing

The city contracted Photomaton Numérique, a French public-sector technology consultancy based in the 13th arrondissement near the Bibliothèque nationale de France site at Tolbiac, to deploy a perceptual hashing and machine-learning deduplication pipeline across four pilot databases. The pilot, which ran from February through May 2026, covered the Seine-Saint-Denis urban regeneration archive and the heritage records for the 3rd and 4th arrondissements — the Marais district, where the density of listed buildings makes photographic documentation especially intensive. Early results from the pilot, presented to the city's digital infrastructure committee in June, showed a 28 percent reduction in storage overhead and a measurable improvement in retrieval accuracy for planners querying the Marais conservation zone files.

Critically, the system flags rather than auto-deletes. Human archivists at the Médiathèque de l'Architecture et du Patrimoine, which holds national-level records in parallel with city-level archives, retain final authority over what gets removed. That human-in-the-loop design is not universal. It distinguishes Paris from Amsterdam, where the Gemeente Amsterdam digital records office adopted a more aggressive automated purge model in 2024 — and subsequently faced complaints from heritage researchers who found gaps in the photographic record for several canal-district properties.

How Paris Compares to London and Berlin

London's situation is fragmented by design. The Greater London Authority does not maintain a single unified image repository; instead, individual borough councils run their own systems, meaning the duplicate problem exists at 32 separate points of failure simultaneously. A cross-borough working group under the London Digital Taskforce flagged the issue in a March 2026 report, but no consolidated solution has been commissioned. Berlin sits somewhere in between: the Senatsverwaltung für Stadtentwicklung began a deduplication audit in 2023 focused on its Stadtbild database, but the project stalled amid budget negotiations and was only partially resumed this spring.

Paris benefits, in this context, from its relatively centralised governance structure. The Mairie de Paris controls a larger share of its digital infrastructure in-house than most comparable European capitals, which makes city-wide rollout logistically simpler even if politically complex under the current National Assembly pressure on municipal spending. The full expansion of the deduplication program beyond the pilot phase is budgeted at approximately €2.1 million over two years, pending approval from the Paris city council's digital and innovation committee, which is scheduled to vote on the allocation in September 2026.

For residents and professionals who depend on these records — architects submitting planning applications in the Marais, historians cross-referencing Seine-side building surveys, or journalists pulling images from public portals — the practical advice is straightforward for now: verify any archival image sourced from city databases against at least one secondary repository, such as the BnF's Gallica platform, until the full deduplication rollout is confirmed. The pilot has cleaned up a meaningful slice of the problem. The rest of the archive is still a work in progress.