How Los Angeles Ended Up With a Duplicate Image Problem — and Why It Now Has to Fix It

Los Angeles city officials and private contractors are confronting a sprawling, largely invisible crisis hiding inside their own servers: tens of thousands of duplicate images spread across municipal databases, tourism platforms, and 2028 Olympic planning portals that are slowing systems, inflating storage budgets, and — in at least some documented cases — feeding outdated or misleading visuals into public-facing communications.

The problem did not appear overnight. It is the accumulated result of more than a decade of digital fragmentation, during which city departments, the Los Angeles Tourism and Convention Board, the LA28 Olympic organising committee, and entertainment industry vendors all built parallel content libraries with little coordination and no shared deduplication standard.

How the Backlog Built Up

The inflection point arrived somewhere around 2020, when pandemic-era remote work forced a chaotic migration of assets onto cloud platforms. The City of Los Angeles Information Technology Agency, headquartered downtown on Spring Street, documented a sharp spike in unstructured file storage between 2020 and 2022 as departments uploaded working files, press photos, and archive images to platforms including Microsoft SharePoint and Google Workspace — often without retiring the originals from legacy servers. By 2023, the city's IT budget line for cloud storage had become one of the fastest-growing line items in the general technology fund, according to budget documents reviewed during prior fiscal cycles.

The entertainment sector compounded the issue. After the Writers Guild of America strike of 2023 and the subsequent AI disruption to production workflows, studios and streaming platforms began bulk-importing stills, frame grabs, and promotional assets into machine-learning training libraries. Companies operating out of the Culver City media corridor and along the Sunset Strip uploaded enormous image sets to content management systems that were never designed for deduplication at scale. The result was redundant libraries that now sit, unresolved, across production pipelines.

In South Los Angeles, nonprofit organisations administering Mayor Karen Bass's Inside Safe homelessness housing program ran into a version of the same problem at street level. Outreach coordinators using tablet-based intake software found that duplicate profile photos — sometimes created when clients enrolled at different service sites, including the L.A. Grand Hotel shelter downtown and facilities in the Westlake neighbourhood — were generating duplicate case records. Social workers flagged the issue to program administrators, though a formal audit timeline has not been publicly announced.

The Stakes Heading Into 2028

With the 2028 Summer Olympics now less than two years away, the consequences of leaving duplicate image infrastructure unresolved are sharpening. LA28 is building out a digital asset management platform intended to serve accredited media from roughly 200 countries. Engineers contracted to the project have described a baseline requirement: every photo asset must carry a unique identifier to prevent duplication errors during broadcast distribution. The Exposition Park campus — home to the Los Angeles Memorial Coliseum and the future Olympic swimming venue — is among the sites generating the highest volume of architectural and event-space photography, much of which currently exists in multiple versions across different vendor drives.

The cost of inaction is concrete. Commercial cloud storage rates on platforms such as AWS S3 currently run in the range of $0.023 per gigabyte per month for standard tier storage, a figure that scales quickly when libraries contain thousands of near-identical image files. For a mid-sized city agency managing a 50-terabyte archive — not an unusual figure for a department like the Los Angeles Department of Cultural Affairs — unreduced duplication can translate to tens of thousands of dollars in avoidable annual costs.

Several paths forward are already in use elsewhere in the public sector. The Los Angeles County Department of Arts and Culture began piloting a perceptual hash deduplication process in late 2024, applying it first to the publicly accessible LA County Image Archive. The approach uses algorithmic fingerprinting to identify visually identical or near-identical images regardless of file name or metadata. City IT officials are expected to present recommendations to the Mayor's office this autumn, with any new procurement tied to existing Digital Equity and Access Initiative guidelines. Organisations managing their own content libraries — from community groups in Boyle Heights to production companies in Hollywood — can start the process now by auditing file-naming conventions and centralising uploads through a single content management point before external mandates arrive.