Paris Takes Aim at Duplicate Images Online — But Lags Behind London and Berlin

Paris city hall has acknowledged a systemic problem that cultural institutions across the capital have been quietly wrestling with for two years: the mass duplication of official digital images — photographs, archival scans, heritage records — across public databases, tourism platforms and administrative portals. The issue came into sharp focus this spring after the Bibliothèque nationale de France flagged hundreds of duplicated high-resolution scans in its Gallica platform, where the same archival images were appearing under different catalogue numbers, sometimes with conflicting attribution metadata. Nobody disputes the images are real. The problem is they are everywhere, simultaneously, under different identities.

The timing matters. France's transposition of the EU's Digital Services Act created new accountability obligations for public bodies that host user-facing content. Institutions that run searchable image databases — the BnF, the Musée d'Orsay, and Paris Musées, the network overseeing fourteen of the city's municipal museums — now face audit requirements tied to content integrity. Duplicate records are not merely an archival nuisance; under the evolving framework, they can constitute misleading metadata, which carries compliance risk. The Grand Paris Express construction project has added pressure: dozens of newly commissioned documentary photographs of tunnel works and station sites around La Défense and Saint-Denis have already appeared duplicated across contractor portals, city council publications and infrastructure press releases, each copy carrying slightly different captions.

What Paris Is Actually Doing About It

Paris Musées launched an internal deduplication audit in January 2026, covering its shared collections management system across venues including the Musée Carnavalet on Rue des Francs-Bourgeois and the Petit Palais on Avenue Winston-Churchill. The audit, described in a February board document circulated to city councillors, identified more than 12,000 image records flagged for manual review out of a digitised collection of roughly 320,000 items. Progress has been slow. As of late June, fewer than a third of flagged records had been resolved, according to the same document. Staff shortages and an underpowered deduplication algorithm — one that relies on filename matching rather than perceptual hashing — have been cited internally as the main bottlenecks.

The BnF, operating under the Culture Ministry rather than the city, is further along. It adopted a perceptual hashing protocol for Gallica in late 2024, a method that compares the visual content of images pixel-by-pixel rather than relying on metadata labels. That system has cleared roughly 40,000 duplicate entries since deployment, though the BnF has not published a full accounting of how many remain.

How Paris Compares to London and Berlin

London's approach has been more centralised. The Victoria and Albert Museum and the British Museum both integrated deduplication tooling into their Collections Online platforms in 2023, following a joint recommendation from the UK's National Archives. The British Museum's internal figures, published in its 2024-25 annual report, cited a 94 percent reduction in duplicate image records over eighteen months. That is a benchmark Paris has not come close to matching.

Berlin moved even faster after Germany's Federal Archive mandated a unified metadata standard for all publicly funded cultural databases in January 2025. The Staatliche Museen zu Berlin — which oversees seventeen collections — completed a system-wide deduplication sweep by October 2025. The effort cost approximately €2.3 million and was funded partly through a federal digitisation grant under the Digitale-Zukunft programme.

Paris has no equivalent central mandate. The city's digital infrastructure directorate, the DSTI, published a roadmap in March 2026 outlining ambitions for shared image-identity standards across municipal bodies, with a target implementation date of mid-2027. Critics within the cultural sector argue that date is optimistic given current staffing and budget allocations. France's broader digital governance budget, cut by roughly 8 percent in the 2026 national budget passed under Assembly pressure, has left institutions competing for shrinking technical resources.

For Parisians who use platforms like Gallica or Paris Musées's online collections — whether for academic research, journalism or simple curiosity — the practical advice is straightforward: cross-reference any image against at least one additional institutional source before citing its metadata as authoritative. The catalogue number alone is no longer a reliable guarantee of uniqueness. Until the deduplication audits run their course, the city's digital archives remain a work in progress, and a patchier one than officials would prefer to admit.