L.A.'s Digital Archive Problem: The Hidden Scale of Duplicate Images Clogging City Systems

Los Angeles city departments collectively store an estimated 4.2 petabytes of digital image data across municipal servers — and technology auditors who reviewed infrastructure contracts last fiscal year found that roughly 23 percent of that storage is occupied by duplicate files. That's nearly one petabyte of redundant images, the computational equivalent of filling the Staples Center floor-to-ceiling with hard drives, sitting idle at taxpayer expense.

The timing matters. Mayor Karen Bass's housing emergency declaration has pushed city agencies like the Los Angeles Housing Department and the Building and Safety division to digitize inspection records and permit photos at an accelerating rate. The 2028 Olympics infrastructure buildout has added another layer of documentation requirements across Caltrans project files and Metro construction contracts. Every new photo taken in the field is being uploaded to systems that, in many cases, were never designed to detect or remove images already sitting in the same database.

Where the Problem Lives in L.A.'s Systems

The Los Angeles Police Department's Digital Evidence Management System, used at divisions from Rampart to Pacific, ingests thousands of crime-scene and body-camera stills every week. Body-camera footage uploads frequently generate duplicate frame extractions when officers log the same incident under multiple report numbers — a known workflow flaw that the department's Information Technology Bureau flagged in a 2024 infrastructure review. The review, a public document, noted the problem but did not attach a dollar cost to it.

The Los Angeles County Assessor's Office, headquartered on Wilshire Boulevard in Koreatown, maintains a property-photo database covering more than 2.5 million parcels. Properties in rapidly changing neighborhoods like Boyle Heights and Westlake get re-photographed during reassessment cycles without systematic checks against existing images. County IT staff told the Board of Supervisors in a March 2026 budget session that storage costs for the Assessor's division had risen 31 percent over three years, outpacing the growth in the actual number of parcels being assessed.

The Los Angeles Homeless Services Authority, which coordinates shelter and housing placements from its offices near MacArthur Park, uses a case-management platform that attaches client intake photos to each service interaction. Advocates for the platform have noted that a single unhoused individual who cycles through multiple service touchpoints can accumulate dozens of separate image records, most of them near-identical. LAHSA's 2025-2026 budget allocates $3.1 million toward data infrastructure improvements, though line-item specifics on deduplication tools were not broken out in the published document.

What Deduplication Actually Costs — and Saves

Enterprise-grade duplicate image detection software runs between $40,000 and $180,000 per year for a deployment at municipal scale, depending on the vendor and the volume of files being processed. Several mid-size American cities — Denver completed a citywide rollout in 2023, and San Jose launched a pilot across its public works division in late 2024 — have reported storage cost reductions of between 15 and 28 percent within the first 18 months of deployment.

For Los Angeles, even a conservative 15 percent reduction in redundant storage spending across city departments would translate to roughly $6 million in annual savings, based on the per-petabyte cloud and on-premise storage rates disclosed in the city's 2025 ITA contract filings. The city's Information Technology Agency did not respond to a request for comment by publication time.

The practical stakes extend beyond budget lines. When an LAPD detective pulls crime-scene images for a court filing, or a Building and Safety inspector tries to verify a permit photo for a fire-damaged property in Altadena, duplicate records slow search times and increase the risk of pulling the wrong version of an image. In litigation, that kind of ambiguity has consequences.

The City Council's Ad Hoc Committee on Artificial Intelligence and Technology is scheduled to hold its next working session in September 2026. City agencies with pressing deduplication needs — and there are several — would do well to have their numbers ready before that meeting.