San Francisco Leads Cities in Purging Duplicate Property Photos — But the Gap With London and Singapore Is Closing Fast

San Francisco's Department of Technology has quietly made duplicate image removal a priority in the city's property listing infrastructure, a push that accelerated after an internal audit found thousands of recycled photographs across active listings on the city's affordable housing portal, DAHLIA, operated by the Mayor's Office of Housing and Community Development. The audit, completed in spring 2026, flagged the problem as a direct obstacle to housing transparency at a moment when the city faces a court-enforced mandate to produce 82,069 new units by 2031 under its state-assigned Regional Housing Needs Allocation.

The practical stakes are higher than they sound. Duplicate images — the same photograph of a kitchen or bedroom appearing across multiple listings, sometimes for properties blocks apart — distort how renters assess the market. A Mission District renter comparing units on Shotwell Street with units near 16th Street BART may be looking at identical stock photography without knowing it. That erodes trust in the city's affordable housing pipeline at the worst possible time.

How SF's Cleanup Compares to London, Singapore, and Toronto

San Francisco's approach centers on perceptual hash matching, a technique that detects visually near-identical images even when file names or metadata differ. The Department of Technology began piloting the system on DAHLIA in January 2026, and by June the office reported that the portal had cleared more than 4,000 flagged duplicates from active listings. Private platforms operating in the city — including Zillow and Apartments.com — are not subject to the city's audit authority, though the Department of Technology has opened informal discussions with both companies about voluntary compliance.

London's approach, run through the Greater London Authority's digital infrastructure team, is further along. The GLA integrated automated image deduplication directly into its own housing data feeds in late 2024, covering roughly 340 borough-level listing databases. Singapore's Housing and Development Board, which controls the majority of residential real estate on the island, has used machine-learning image validation since 2023 as part of a broader digital registry overhaul. Toronto's city government mandated image uniqueness requirements for listings appearing on its Affordable Rental Housing portal starting March 2025, with financial penalties for non-compliant landlords reaching CAD $500 per violation.

San Francisco has no penalty mechanism yet. The DAHLIA cleanup is voluntary for property managers, and the Department of Technology's enforcement toolkit remains limited to delisting flagged units — a measure housing advocates in SoMa and the Tenderloin have called insufficient given the scale of the problem.

What Comes Next for Renters and Property Managers

The Mayor's Office of Housing and Community Development is expected to circulate a draft policy by September 2026 that would require image uniqueness certification for all listings appearing on city-affiliated portals. Whether that extends to privately operated platforms will depend on state-level authority the city does not currently hold. Legislation introduced in Sacramento earlier this year — AB 2214, which targets deceptive practices in residential listings — could give municipalities including San Francisco the legal hook they need, but the bill remained in the Assembly Appropriations Committee as of June 30.

For renters navigating the market right now, the most reliable approach is to cross-reference listing photographs against Google Street View for exterior shots and to request a live video walkthrough before signing any lease. The Tenderloin Housing Clinic on Turk Street, which serves low-income renters across several neighborhoods, has begun advising clients to flag suspected duplicate image use to the city's 311 system, which now routes those complaints to the Department of Technology's new digital housing integrity team.

San Francisco's effort is real, but it is still catching up. London and Singapore built deduplication into their housing data architecture from the start. SF is retrofitting a system that was never designed to catch the problem, and the city's private rental market — where most renters actually search — remains almost entirely outside the city's reach.