Los Angeles city archivists and planning officials have been quietly wrestling for more than a year with a problem that sounds mundane until you see the price tag: tens of thousands of duplicate or mismatched photographs embedded in public-facing permit databases, property records, and homelessness service intake systems. With the 2028 Olympics infrastructure buildout accelerating and Mayor Karen Bass's housing emergency reshaping how the city processes construction permits, the stakes for clean, accurate digital records have never been higher.
The city's Bureau of Engineering and the Department of City Planning together manage digital record systems that span roughly 1.2 million active parcels across Los Angeles County. When the same site photograph — or worse, a misidentified image of the wrong property — appears attached to multiple permit applications, it creates legal exposure, slows project approvals, and can misdirect homeless outreach workers trying to document encampment conditions before a cleanup. The problem surfaced prominently in 2024 during audits of the Comprehensive Homeless Strategy database, and officials have been developing remedies ever since.
What LA Is Actually Doing About It
The city's Information Technology Agency, based downtown on Main Street, has been piloting a deduplication protocol using hash-matching software — a method that assigns each image a unique digital fingerprint and flags identical files across databases. The pilot launched formally in January 2026 and initially covers records processed through the Los Angeles Homeless Services Authority, known as LAHSA, and the Planning Department's development portal on Spring Street. According to city budget documents reviewed for fiscal year 2025-26, the ITA allocated approximately $2.3 million toward data integrity projects across multiple departments, though that figure encompasses broader database work beyond images alone.
The Echo Park and Boyle Heights neighborhoods have served as early test zones, partly because both areas generated high volumes of field photography during encampment documentation between 2022 and 2025. Duplicate images in those records had, in some cases, caused the same location to appear as two separate service entries, skewing LAHSA's outreach count data. The ITA's deduplication sweep of those two neighborhoods' records alone identified more than 4,000 flagged image pairs by March 2026, city technology briefing documents show.
London and Seoul Are Further Along — but With Different Problems
Other major cities have been grappling with the same issue, with mixed results. Transport for London began a systematic image deduplication project across its congestion charge camera network and planning portal in 2023, contracting with a UK-based data management firm after an internal review found duplicate entries were inflating infrastructure damage assessments following the 2021 utility upgrade cycle. London's approach relied on a centralized image registry — a single repository that all city agencies must upload to before attaching photos to any public record. That model has reduced reported duplicate rates in TfL planning documents by a margin cited in a 2025 Greater London Authority oversight report, though the registry required a reported £4.7 million build-out cost.
Seoul's city government took a different path. Beginning in 2022, the Seoul Digital Foundation embedded AI-based image classification directly into the city's smart city platform, which processes construction permit imagery in near real time. The system flags probable duplicates within seconds of upload. South Korean press accounts from 2025 reported that the foundation estimated a reduction of manual review labor by the equivalent of dozens of full-time positions annually, though the foundation's English-language materials do not specify the exact count.
Los Angeles is, by most independent assessments of city technology spending, working with a smaller per-capita digital infrastructure budget than either London or Seoul. The ITA's hash-matching pilot is pragmatic rather than cutting-edge — it does not yet use machine-learning classification, which means visually similar but not identical images still slip through. A broader AI-assisted phase has been discussed internally but has not been formally funded as of this reporting.
For residents and contractors, the practical advice is straightforward: anyone submitting permit applications through the city's Development Services Center on Figueroa Street should upload photographs with embedded GPS metadata and standardized file-naming conventions, both of which the ITA says dramatically reduce the likelihood of a misfiled or duplicated image slowing a project's review. The city's online portal updated its submission guidelines on this point in April 2026. With Olympic venue construction timelines tightening, getting the paperwork right the first time is no longer just administrative courtesy — it is a deadline.