How Paris's Duplicate Image Problem Crept Into the City's Digital Public Record — And Why Fixing It Now Matters

Paris city hall confirmed this spring that its centralised digital image repository, managed through the Direction des Affaires Culturelles, contains an estimated 40,000 duplicate or near-duplicate image files accumulated since the first major digitisation push began in 2011. The problem did not arrive overnight. It is the product of more than a decade of competing institutional priorities, budget cycles and a sprint of visual documentation tied to the 2024 Summer Olympics that flooded municipal servers with uncatalogued material.

The stakes are practical as much as archival. As the Mairie de Paris accelerates its open-data commitments and the Grand Paris Express project demands a steady pipeline of public-facing communications imagery — construction milestones, community consultation visuals, progress maps — duplicates clog search systems, slow editorial teams and, in several documented cases, sent conflicting versions of the same landmark photograph into circulation across official channels simultaneously.

A Timeline of How the Backlog Accumulated

The roots go back to the early 2010s, when the city launched parallel digitisation contracts across different administrative departments. The Bibliothèque historique de la Ville de Paris on rue de Rivoli ran its own scan-and-catalogue operation. The Pavillon de l'Arsenal, the city's urbanism documentation centre on boulevard Morland in the 4th arrondissement, maintained a separate image bank for planning and architecture records. Neither system spoke to the other in any automated way.

By 2018, the problem was already visible internally. A cross-departmental audit that year — referenced in public budget documents submitted to the Conseil de Paris — flagged interoperability failures between at least six distinct content management systems operating under the city umbrella. Recommendations were issued. Funding was partially allocated. The merger work stalled.

Then came Paris 2024. The preparation phase alone generated an extraordinary volume of photographic and video documentation: construction at Saint-Denis, venue fit-outs at Le Grand Palais, Seine-Saint-Denis community engagement sessions. Accredited photographers, city communications officers and partner agencies all uploaded material, often to different platforms with inconsistent metadata standards. The result, according to the spring 2026 confirmation from the Direction des Affaires Culturelles, was a post-Games backlog that compounded every pre-existing flaw in the system.

The Clean-Up and What Comes Next

The current remediation project, budgeted at roughly €1.2 million over eighteen months according to documents presented to the city council's culture and heritage committee in March 2026, centres on deploying perceptual hashing software capable of identifying visually identical or near-identical images regardless of file name or format. The contract was awarded to a consortium including a French digital heritage firm, with implementation scheduled to run through the end of 2027.

The work is also tied to a broader ambition: feeding a cleaned, properly tagged image library into the open-data portal at opendata.paris.fr, which already hosts datasets used by journalists, researchers and urban planning consultancies tracking Seine riverbank regeneration and banlieue infrastructure change. Duplicate removal is, in that sense, a prerequisite for credibility. A dataset that points to six versions of the same 2019 photograph of the Porte de la Chapelle arena construction is not a useful public resource.

For Parisians and institutions that rely on the city's image commons — from neighbourhood associations in the 18th arrondissement preparing local history exhibitions to architects pulling reference material for Grand Paris Express station design reviews — the practical advice is simple: treat any image downloaded from official city platforms before mid-2027 as potentially part of a duplicated set, cross-check metadata and, where precision matters, request the canonical version directly from the Direction des Affaires Culturelles. The clean library, officials have indicated, will be fully public-facing once validation is complete. For now, the work of sorting fifteen years of well-intentioned but poorly coordinated documentation is grinding forward, one file at a time.