How Los Angeles Is Tackling Duplicate Images in Its Public Records — and How It Stacks Up Against Cities Worldwide

Los Angeles city departments are sitting on millions of duplicate digital images — redundant photos, scanned documents stored twice, and mislabeled files clogging servers across agencies from the Department of Public Works to the Los Angeles Homeless Services Authority. The problem, long dismissed as a low-priority IT headache, has taken on new urgency as the city accelerates its digitization push ahead of the 2028 Summer Olympics and under pressure from Mayor Karen Bass's ongoing housing emergency declaration, which has generated a paper and photo trail spanning thousands of properties.

Duplicate image replacement — the process of identifying, reconciling, and replacing redundant visual files in public databases — sounds unglamorous. But for a city that spent roughly $47 million on its digital infrastructure overhaul in fiscal year 2025, according to the city's published budget documents, the inefficiency carries a real price tag. Bloated image libraries slow down planning portals, complicate public records requests, and create legal exposure when contradictory photos of the same property appear in different agency files.

What LA Is Actually Doing About It

The Bureau of Engineering, headquartered downtown on South Figueroa Street, began piloting a deduplication protocol in early 2025 as part of its GeoHub platform expansion. The GeoHub — LA's open data mapping system — hosts aerial imagery, street-level photos, and infrastructure documentation that is accessed by contractors, planners, and journalists. Engineers identified thousands of image pairs where the same parcel had been photographed under different file names and stored separately, sometimes years apart.

The Los Angeles County Office of Digital Services has separately been coordinating with the city's Information Technology Agency on a broader cleanup, particularly affecting images tied to the Skid Row and Boyle Heights census of unhoused encampments conducted in 2024. Those images, captured for compliance purposes under the Bass administration's Inside Safe program, were in some cases uploaded by multiple field workers on the same day, creating near-identical duplicates that consumed server space and made audit trails harder to follow.

The city's approach leans heavily on hash-matching software — a technique that assigns a unique numerical fingerprint to each image file — rather than AI-based visual recognition tools, which LA's ITA has been cautious about deploying at scale given the ongoing debate at City Hall over algorithmic accountability. That caution has a cost: hash-matching catches exact duplicates but misses near-identical images taken seconds apart, which is where much of the redundancy actually lives.

How Other Major Cities Are Handling the Same Problem

London's Government Digital Service began rolling out AI-assisted deduplication across borough council archives in late 2024, using visual similarity scoring to catch near-matches that exact-fingerprint tools miss. The approach is further ahead than anything LA has publicly committed to. Amsterdam's municipal GIS team, working with Kadaster — the Dutch national land registry — resolved more than 200,000 duplicate cadastral images between 2023 and 2025, a figure the city published in its annual digital governance report. Tokyo's ward offices, which manage property and infrastructure photography locally, standardized a single upload portal in 2022 that rejects duplicate submissions at the point of entry, preventing the problem before it metastasizes.

LA's decentralized agency structure works against that kind of upstream fix. Unlike Tokyo's ward system or Amsterdam's unified land registry, Los Angeles has more than 40 city departments that each manage their own image storage, often with different software vendors and different retention schedules. Getting them to talk to each other — let alone share a single upload portal — has proven to be a years-long coordination challenge that no single office has authority to force.

For residents or attorneys trying to pull property records from the city's ZIMAS planning portal or submitting public records requests to the City Clerk's office on North Spring Street, the practical advice is straightforward: request that any image files be flagged with the date of capture and the uploading department, which helps identify which version is authoritative when duplicates appear. The ITA has confirmed that staff can fulfill requests in this format, though it is not yet the default.

The next benchmark arrives in early 2027, when the Olympic planning authority LAOOC will require fully rationalized digital asset libraries from all participating city agencies. That deadline, more than any internal IT directive, may finally force the consolidation that years of budget presentations and interagency memos have not.