Paris Moves to Purge Duplicate Images From Its Public Archive: The Key Decisions Ahead

Paris's public image archive—spanning tens of thousands of photographs held by the Direction des Affaires Culturelles de Paris and distributed across municipal portals—contains a substantial volume of duplicate files that distort search results, bloat server costs and create legal uncertainty over which version of an image carries the authoritative rights clearance. The city has now confirmed that a formal deduplication programme will begin before the end of 2026, placing a cluster of expensive and politically sensitive decisions directly in front of administrators who are already stretched by Grand Paris Express construction demands and post-Olympic legacy commitments.

The timing matters. Paris spent heavily on digital infrastructure for the 2024 Olympics, including a consolidated media asset management system piloted at the Stade de France and rolled out to satellite venues across Seine-Saint-Denis. That system left behind a fragmented inheritance: image libraries from dozens of press pools, partner agencies and official city channels that were never fully reconciled. Estimates circulating inside the Hôtel de Ville—though not yet published in any official budget document—suggest the duplicate problem affects a meaningful share of the roughly 400,000 assets now sitting in the municipal digital repository. The city has not released a precise figure.

What the Deduplication Process Actually Involves

The core technical challenge is not simply deleting identical files. Many duplicates differ by resolution, cropping, watermark or metadata tag, meaning automated hash-matching tools flag them as distinct assets even when they depict the same moment. The city's information systems directorate, the Direction des Systèmes d'Information et du Numérique, is understood to be evaluating at least three vendor proposals for semi-automated deduplication software. A decision on procurement is expected in the autumn, ahead of the 2027 municipal budget cycle.

Two institutions sit at the centre of the coming choices. The Bibliothèque Historique de la Ville de Paris, on rue Pavée in the Marais, holds digitised historical collections that overlap substantially with images also stored on the Paris Musées platform. Curators there have long flagged the risk that aggressive automated deduplication could strip contextual metadata from archival prints—information about provenance, donor and conservation state that is attached to what a machine might read as a redundant copy. The Paris Musées consortium, which coordinates the digital offer of fourteen city-owned museums including the Musée Carnavalet on rue de Sévigné, has its own governance board and its own interpretation of which file constitutes the master record.

Getting those two institutions to agree on a shared deduplication protocol is the political problem underneath the technical one. The Direction des Affaires Culturelles will need to arbitrate, and it is doing so under budget pressure. The 2026 city budget allocated roughly €4.2 billion across all municipal services, with cultural affairs receiving a share that administrators have described publicly as flat in real terms relative to 2025.

The Decisions That Cannot Wait

Three questions will define what the programme looks like in practice. First, who holds the master rights record when two versions of an image exist under different licence agreements—one, say, acquired during a temporary Olympic partnership and another held permanently by the city? Second, how does the city handle images where the duplicate is the only surviving high-resolution copy, even if it was ingested twice by accident? Third, what public-access guarantees apply once files are deleted, given that several neighbourhood associations in arrondissements including the 19th and the 20th have begun using the municipal archive for local history projects tied to the Grand Paris Express corridor regeneration.

A working group is expected to report to the cultural affairs cabinet in September. If the city approves a vendor contract before year-end, the active deduplication phase could begin in the first quarter of 2027—though heritage advocates say that timeline is optimistic given the legal review each disputed file category will require. What is not in dispute is that the status quo carries its own cost: server contracts for the city's primary data centre in the 13th arrondissement are up for renewal in mid-2027, and carrying redundant assets into that renewal will inflate the bill. The pressure to act is financial as much as it is archival.