How Paris's Archives Ended Up Full of Duplicates: The Long Road to a Reckoning

Paris city hall confirmed this spring that a coordinated audit of digital image assets held across municipal departments had turned up hundreds of thousands of duplicate files — some stored three and four times over, across servers maintained by separate directorates that had never been asked to talk to each other. The problem did not appear overnight. It is the residue of roughly twenty years of digitisation drives, each launched with urgency and each operating in near-total isolation from the last.

The timing matters. With the Grand Paris Express still expanding — Line 15 South opened in 2023 — and the Paris 2024 Olympics legacy machine now in full activation mode, the city's communications and heritage departments are under pressure to produce coherent visual records faster than ever before. Duplicate imagery clogs retrieval systems, inflates storage costs and, in several documented cases, has meant that rights-cleared photographs were used alongside visually identical images whose licensing status was unknown. That is not an abstract risk: it carries real legal and financial exposure for the Mairie de Paris.

A Fragmented History of Digitisation

The roots of the problem stretch back to the early 2000s, when the Direction des Affaires Culturelles launched its first major push to digitise holdings from institutions including the Bibliothèque Historique de la Ville de Paris, on Rue de Rivoli, and the Musée Carnavalet, on Rue de Sévigné in the Marais. Those two institutions alone digitised tens of thousands of images between 2001 and 2010, according to the city's own published heritage reports. The problem was that each institution used its own metadata schema, its own file-naming convention and, critically, its own server infrastructure.

Then came the 2014 launch of Paris Open Data, the municipal platform designed to make city-held assets publicly accessible. Staff at multiple departments uploaded image sets to the platform independently, without cross-referencing what had already been deposited. The Seine-Saint-Denis prefectural archive, which collaborates with Paris on several Grand Paris projects, ran into the same wall when trying to merge visual records for joint urban regeneration documentation along the Seine riverbanks in 2019. By that point, the duplication problem was systemic, not incidental.

A 2022 internal review, referenced in a published summary by the city's Direction de la Transformation et des Relations avec les Usagers, found that storage costs attributable to redundant files across the municipal network had risen measurably year-on-year since 2015. The review did not publish a precise euro figure, but it characterised the cost trajectory as unsustainable given ongoing budget pressure from the National Assembly over municipal subsidies.

What the Clean-Up Actually Involves

Replacing or consolidating duplicate images is not simply a matter of deleting files. Each image in a public archive may have multiple rights statuses, multiple use histories and multiple metadata records — all of which have to be reconciled before any file is removed. The city contracted with a specialist digital asset management firm in late 2025 to begin the deduplication work across three priority departments: the Direction du Patrimoine et de l'Architecture, the communications unit overseeing the Hôtel de Ville, and the team managing the Paris 2024 legacy photographic record based out of the Stade de France coordination office in Saint-Denis.

The 2024 Olympics alone generated an estimated archive of several million images, a volume that overwhelmed existing sorting protocols almost immediately after the Games closed in August of that year. Legacy activation — meaning the ongoing public and commercial use of those images to promote Paris as a destination and to document the post-Games urban transformation — depends on a clean, searchable, rights-clear library. Right now, that library does not fully exist.

Administrators say the deduplication programme is expected to run through the end of 2026. For Paris residents and researchers who rely on the open data portal or visit reading rooms at institutions like the Carnavalet or the Archives de Paris on Boulevard Sérurier in the 19th arrondissement, the practical advice is straightforward: flag any apparent duplication when submitting image requests, and expect slower-than-usual turnaround times on rights clearances until the audit concludes.