'My Whole Story Was Gone': Angelenos Speak Out on Duplicate Image Removal Sweeping Local Archives

Families across Los Angeles are discovering that photographs documenting decades of neighborhood life — block parties in Boyle Heights, quinceañeras along Cesar Chavez Avenue, vigils held after the 2025 Eaton fire — have been quietly deleted from digital archives managed by city-contracted platforms. The culprit, according to community advocates and library staff who have reviewed the issue, is aggressive deduplication software that flags visually similar images for removal, then purges them automatically before any human review.

The problem has surfaced with particular urgency in mid-2026 because several Los Angeles agencies are in the middle of large-scale digitization drives tied to the 2028 Olympics infrastructure push and Mayor Karen Bass's broader effort to modernize city services. More images are being uploaded faster than at any point in the city's archival history, and the deduplication algorithms — designed to save server costs — are moving just as fast.

What Gets Lost When the Algorithm Decides

The Los Angeles Public Library's Braxton Collection Management System, used across the Central Library on West Fifth Street and 72 branch locations, runs automated deduplication passes every 72 hours. When two images share more than an 85 percent pixel-similarity score, the system marks one for deletion. Librarians say the threshold is too low to distinguish between a true duplicate and two photographs taken seconds apart — one sharp, one the only surviving image of a specific face or moment.

The East LA Community Corporation, which maintains a digital oral history archive built over 14 years in the Unidad neighborhood near 1st and Lorena streets, says it lost an estimated 340 images in a single automated pass in March after migrating to a new cloud storage platform. The organization said it has been working since April to reconstruct the affected files from physical prints held by community members, a process staff describe as painstaking and still incomplete.

Residents reached through community meetings organized by Self-Help Graphics & Art on Cesar Chavez Avenue have described the deletions in stark personal terms. One woman said photographs of her late mother taken at a 2019 Día de los Muertos celebration were among the files removed. Another described submitting family images to a neighborhood history project only to be told months later that the originals could not be recovered. Neither wanted to be named in print while they continue to seek resolution from the agencies involved.

A Systemic Gap With No Easy Fix

The deduplication issue is not unique to Los Angeles, but the city's scale makes it acute. The Los Angeles City Archives holds more than 1.4 million digital image files, according to figures published in the City Clerk's 2025 annual report. Even a fraction of a percent of erroneous deletions translates to thousands of lost records.

The California State Library distributed a $2.3 million grant in fiscal year 2025-26 to fund community archiving projects statewide, with Los Angeles County institutions receiving the largest share. That money is accelerating uploads without, critics say, requiring grantees to adopt human-review protocols before automated deletions occur.

Advocates are now pressing the city's Department of Cultural Affairs, which oversees several affected grant programs, to require a 30-day human review window before any automated deletion is finalized. The department has not publicly responded to the proposal as of July 4.

For families waiting, the practical advice from archivists is blunt: do not rely on a single digital repository. The Los Angeles County Museum of Art's Boone Children's Gallery archive team recommends residents keep personal copies on at least two separate storage media — an external hard drive and a cloud service not affiliated with the city — before submitting originals to any institutional program. The Echo Park-based nonprofit Dignity in Documentation has published a free one-page guide, available at its Vermont Avenue office, walking community members through a basic three-copy backup process. In the meantime, the March deletions at East LA Community Corporation remain only partially recovered, and summer digitization drives are still running.