DC Archives and Cultural Sites Push to Fix a Duplicate Image Problem That's Been Hiding History

The DC Public Library's Special Collections division confirmed this week that it is conducting a systematic audit of its digitised photo holdings after staff identified hundreds of duplicate image entries that have been mislabelled, buried, or effectively removed from public search results. The problem, which librarians say has compounded over years of rushed digitisation drives, means researchers using the Martin Luther King Jr. Memorial Library on G Street NW may be searching incomplete records without knowing it.

The timing matters. The library's digitisation push accelerated after its 2020 renovation reopening, and several partner institutions — including the DC History Center at the Carnegie Library on Mount Vernon Square — have been cross-referencing collections in preparation for a joint online portal expected to launch later this year. Duplicate image entries create chain errors: when one record is flagged and suppressed, legitimate unique images attached to the same metadata string can disappear from search indexes entirely.

What Went Wrong and When

The root cause appears to be a combination of batch-upload errors and inconsistent file-naming conventions during digitisation sprints between 2021 and 2024. Institutions that contracted external vendors to scan and catalogue physical photographs frequently received deliveries of image files with near-identical metadata — same date stamp, same collection tag — that automated deduplication software then collapsed into single records. Photographs documenting neighbourhood change in Anacostia, Shaw, and NoMa were among the collections flagged in internal reviews as potentially affected, according to library staff communications obtained through a public records request filed by this newspaper.

The DC History Center, which holds one of the city's most comprehensive collections of ward-by-ward neighbourhood photography, began its own internal review in late June after a researcher working on a project about H Street NE's commercial corridor reported finding fewer than 40 searchable images in a subcollection that physical ledgers suggested should contain more than 200. The discrepancy pointed directly to the duplicate-suppression problem. The centre has not yet released a final count of affected records.

The Smithsonian Institution's digitisation office at the National Museum of American History on Constitution Avenue NW has dealt with a version of this problem at scale. The museum's online collections portal, which as of early 2026 listed more than 1.8 million digitised objects, implemented a manual review layer in 2023 specifically to catch automated deduplication errors after similar complaints from academic researchers. That internal fix took roughly 14 months to roll out fully across primary collections.

What Institutions Are Doing Now

The DC Public Library's audit is expected to run through September 2026. Library administrators have not specified how many records are under review, but staff familiar with the project have indicated that the Special Collections photo holdings — roughly 85,000 digitised images in total — will be checked against physical card catalogues and original donor logs to verify uniqueness before any suppressed records are reinstated.

The DC History Center has separately reached out to the Digital Public Library of America, which aggregates records from hundreds of collecting institutions nationwide, to flag that some of its contributed metadata may be unreliable pending the review. That step is significant: records contributed to DPLA appear in searches by universities and archives across the country, meaning the duplication problem in a Northeast DC neighbourhood photo file can affect scholarship far beyond the District.

For residents and researchers trying to access these collections in the meantime, the practical advice from library staff is straightforward. If a digital search on the DC Public Library's online portal returns unexpectedly thin results for a specific neighbourhood or date range, request the physical finding aid directly from the Special Collections desk on the third floor of the MLK Library. Those paper records were not affected by the digitisation errors and can confirm whether more images exist than the database currently shows. The History Center on Mount Vernon Square is similarly directing researchers to its in-person reading room, which is open Tuesday through Saturday, while the audit proceeds.

Both institutions say they expect to publish updated, corrected online records before the end of the calendar year. Whether the joint portal launch will be delayed as a result of the audit has not been officially decided.