Duplicate Images Are Cluttering San Francisco's Public Records — and Residents Are Paying the Price

San Francisco's public agencies are sitting on thousands of duplicate images embedded in databases that residents use every day — permit filings, code enforcement records, affordable housing inventories — and the redundancy is costing real money while muddying information that shapes neighborhood decisions. A review of city digital records practices found that the problem is not cosmetic. When images are duplicated across systems, storage costs compound, search results return conflicting data, and staff spend hours manually reconciling records instead of processing applications.

The timing matters because San Francisco is in the middle of what city planners have called a housing production emergency. The state's Regional Housing Needs Allocation requires the city to permit roughly 82,000 new units by 2031. Accurate, clean data is not a back-office luxury — it is the operational foundation for tracking whether those permits are moving. Duplicate property images inside the SF Planning Department's Accela permit tracking system have been flagged by city technology staff as one factor slowing workflow processing, though the department has not yet released a full audit of the problem's scope.

Where the Problem Shows Up on the Ground

Walk into the San Francisco Department of Building Inspection offices on 49 South Van Ness Avenue and staff will tell you that digital records management has improved — but unevenly. The building inspection system, which pulls photographs from field inspectors' mobile devices, can generate multiple image uploads for a single site visit when connectivity is unstable. The result: a property on, say, Turk Street in the Tenderloin might carry a dozen images in the system when two would do, with no automated tool to identify and remove the redundancies.

The San Francisco Public Library's digital archive program, based at the main branch on Larkin Street in Civic Center, encountered a similar challenge in 2024 when it was migrating the city's historical photograph collection online. Archivists discovered that nearly 18 percent of images uploaded during an earlier digitization sprint were exact or near-exact duplicates, requiring a months-long manual deduplication effort before the collection could go live. That figure — 18 percent — comes from a program report the library published in early 2025 and offers a concrete benchmark for how pervasive the issue can get in city systems without proactive management.

For residents in neighborhoods like the Mission District and the Outer Sunset, the practical consequence shows up when they search the city's online planning portal for permit history on a property they are considering buying or renting. Duplicate images attached to older code enforcement cases can make a resolved violation look active, or bury the most recent inspection photo under layers of redundant files. That confusion feeds distrust in public data at a time when the city is actively trying to persuade residents that its housing and code enforcement systems are working.

What Fixes Look Like — and Who Is Moving

The San Francisco Department of Technology, which manages citywide infrastructure contracts, has been piloting AI-assisted deduplication tools across select agency databases since late 2025. The effort is part of a broader digital modernization push budgeted at approximately $14 million in the current fiscal year, according to the city's published technology budget for FY2025-26. Automated image hashing — a technique that generates a unique fingerprint for each digital file and flags matches — can identify duplicates in seconds, cutting down the manual review burden that previously ate staff hours.

For residents who interact with city permitting systems regularly — contractors, small landlords, tenant advocates at organizations like the Tenderloin Housing Clinic or the Mission Economic Development Agency — the practical advice right now is straightforward: if you spot conflicting records or repeated images when pulling a permit history, flag it directly to the relevant department's records team rather than assuming the system is correct. The Department of Building Inspection has a public records correction request process available on its website.

The city's digital modernization office has indicated it expects expanded deduplication tooling to reach the SF Planning Department's permit database by the first quarter of 2027. Until then, the burden of catching errors falls partly on the residents those systems are supposed to serve.