Paris Confronts Its Duplicate-Image Problem: The Key Decisions Ahead

Paris city hall has a filing problem it can no longer ignore. Across the municipal digital estate — from the open-data portal opendata.paris.fr to the cultural archive systems maintained by the Direction des Affaires Culturelles — technicians have identified tens of thousands of duplicate image files clogging databases, distorting search results and, in several cases, causing the wrong photographs to surface on official public-facing platforms. The question now is who decides what gets deleted, who pays for the audit, and how quickly it has to happen.

The timing is not arbitrary. Paris is in the middle of a sustained effort to monetise and communicate the legacy of the 2024 Olympics, and much of that work depends on clean, correctly attributed imagery. The Comité Régional du Tourisme Paris Île-de-France has been pushing partner institutions to refresh their digital assets for a 2026 promotional campaign targeting North American and Asian visitors. Duplicate or mis-tagged images — a photograph of the Trocadéro labelled as the Stade de France, or a construction shot from the Grand Paris Express works filed under Seine riverbank regeneration — undermine that effort at a basic operational level.

Where the Backlog Is Worst

Three institutions are understood to be carrying the heaviest load. The Bibliothèque historique de la Ville de Paris, on Rue des Francs-Bourgeois in the Marais, digitised roughly 180,000 images over a four-year programme that concluded in 2023. Internal reviews since then have flagged duplication rates in certain collections running as high as one file in five. The Paris Musées network, which pools digital assets for fourteen municipal museums including the Musée Carnavalet and the Petit Palais, has been working under a unified content management system since 2021 but inherited legacy catalogues from each institution that were never fully reconciled. And the urban-planning directorate, the Direction de l'Urbanisme, holds thousands of site photographs from Seine-Saint-Denis and the inner banlieue linked to Grand Paris Express station construction — images filed under overlapping project codes that were reassigned when budget lines shifted between 2019 and 2022.

The practical consequences are not theoretical. When the city's housing office used its image database to compile a report on rental-market tensions in the 11th arrondissement last autumn, at least three photographs depicting properties in the 20th were included in error. The mistake was caught before publication, but only because a staff member recognised the street. Automated deduplication tools alone, officials acknowledge, cannot catch geographic mis-tagging of that kind.

What the Next Decisions Look Like

Two broad options are on the table. The first is a centrally commissioned audit — estimated by procurement documents circulating within the Direction des Systèmes et Technologies de l'Information at somewhere between €400,000 and €650,000 for a full-scope contract — that would standardise metadata across all municipal databases and build a shared taxonomy before the end of 2027. The second option is a phased, institution-by-institution clean-up, cheaper in any single budget year but slower and prone to leaving the cross-database inconsistencies intact.

The Conseil de Paris is expected to take up the question in its autumn session, likely in October. Before that, the city's digital services directorate has asked each major holding institution to submit a self-assessment of its duplication exposure by September 1. That document will shape the budget request that lands on councillors' desks.

For smaller cultural venues and neighbourhood mairies already squeezed by the housing pressures in areas like Belleville and the Canal Saint-Martin corridor, the outcome matters practically: they rely on the central asset pool for everything from event promotion to planning consultations. A failed or delayed resolution means those institutions continue working around broken search results, pulling images manually and burning staff time that most do not have. The autumn session vote will not fix the files. But it will determine whether Paris treats this as a one-time investment or an ongoing irritant — and that choice will ripple through every digital project the city touches for the next decade.