How Paris's Duplicate Image Problem Crept Into the City's Digital Infrastructure — and Why It's Finally Being Fixed

Paris's city government is moving to overhaul the digital image archives underpinning everything from tourism portals to planning applications, after internal audits revealed that duplicate photographs and graphics had multiplied unchecked across municipal servers for more than a decade. The consolidation effort, now being coordinated through the Direction de la Transformation Numérique at the Hôtel de Ville, targets a sprawling patchwork of overlapping databases that grew without central oversight as each arrondissement and agency uploaded its own material independently.

The timing matters. Since the Paris 2024 Olympics, the city has pushed aggressively to monetise its digital assets — selling licensed image packs to travel platforms, embedding high-resolution visuals into the Grand Paris Express communication campaigns, and supplying photographs to the Seine-Saint-Denis urban regeneration project. Redundant files drive up licensing confusion and inflate cloud storage invoices that ultimately land on the municipal budget. Worse, duplicate images with conflicting metadata have reportedly caused mismatches on the Paris.fr planning portal, where property developers and individual applicants alike submit and retrieve documentation.

How the Problem Accumulated

The roots stretch back to the early 2010s, when the Mairie de Paris first digitised its photographic archive at the Bibliothèque Historique de la Ville de Paris on Rue de Sévigné in the 4th arrondissement. At the time, each directorate — roads, parks, social housing, culture — maintained its own folder structure on separate servers. There was no shared taxonomy and no deduplication protocol. By the mid-2010s, the city's communications team at the Préfecture de Paris had begun uploading event photography to a parallel system entirely. The Olympics preparation between 2020 and 2024 added another layer: three separate contractor pipelines delivered images of venue construction at sites including the Stade de France, the Bercy Arena, and the temporary structures along the Trocadéro, each tagged under different naming conventions.

A 2023 internal report — cited in subsequent budget documents presented to the Conseil de Paris — flagged that storage redundancy across municipal systems had grown by roughly 340 percent since 2015, with image files accounting for the single largest category of duplicated data. Cloud storage costs for the city's digital estate had risen substantially over the same period, though the full figure remains within documents not yet made public. What is on record: the Conseil de Paris approved a line item of €2.3 million in its 2025 digital infrastructure budget specifically allocated to data rationalisation, with image deduplication named as a primary workstream.

What Comes Next for City Systems

The practical work is now underway at the city's data centre operations managed through Infranum, the municipal digital services arm. Technicians are running automated hash-comparison tools across the legacy archive, flagging files that are identical or near-identical before human reviewers make final deletion decisions. The Bibliothèque Historique on Rue de Sévigné is serving as the canonical reference point: images already catalogued there under the city's official collection take precedence, with duplicates elsewhere marked for removal.

For residents and professionals who use city platforms day-to-day, the most visible change should arrive in the autumn on the Paris Urbanisme portal, where architects and developers searching for reference photographs of listed facades in Montmartre, the Marais, or along the Boulevard Haussmann have long encountered the same image appearing under multiple search results with conflicting dates and credits. The deduplication pass is expected to reduce that archive's effective size by roughly a third, according to technical documentation circulating among contractors, though no official figure has been confirmed publicly.

The broader lesson the city appears to be drawing is structural rather than technical. Going forward, any new image upload to municipal systems will require a unique identifier tied to the department, the date, and the project code — a workflow modelled partly on the system already used by the Institut national de l'audiovisuel for broadcast archives. Whether the new protocol holds across the dozens of semi-autonomous agencies that feed Paris's digital ecosystem will depend on enforcement, not software. That political question is one city councillors are expected to debate when the digital transformation directorate presents its progress report to the Conseil de Paris this September.