Singapore's War on Duplicate and Fake Images: What Officials, Experts and Industry Figures Are Now Saying

Singapore's Infocomm Media Development Authority formally flagged the proliferation of duplicate and manipulated images in digital public services as a priority enforcement concern this year, placing the issue squarely on the agenda of agencies that run everything from citizen-facing government portals to the national property resale marketplace. The timing is not accidental. Generative AI tools have made it trivially cheap to clone, retouch or fabricate visual content at scale, and administrators of high-trust platforms say the volume of flagged duplicate images has grown sharply since 2024.

The stakes in Singapore are particular. HDB's official resale portal, used by tens of thousands of flat buyers annually, relies on property photographs to anchor listings. When the same image appears across multiple listings — sometimes for units in entirely different blocks — it erodes confidence in a market where a four-room flat in Queenstown or Toa Payoh can change hands for well above S$700,000. Officials at the Housing and Development Board have publicly acknowledged the need for stronger image verification without specifying enforcement figures, and the Urban Redevelopment Authority has separately noted the importance of accurate visual documentation in its licensing requirements for property agents.

What the Experts Are Saying

Academics at the National University of Singapore's School of Computing have been working on perceptual hashing and reverse-image detection pipelines that can flag near-duplicate photographs even when they have been cropped, colour-shifted or slightly reframed — techniques commonly used to defeat basic hash-matching filters. Researchers there have pointed out that the challenge is not purely technical. Platforms must decide what threshold of visual similarity constitutes a policy violation, and that is ultimately a governance question as much as an algorithmic one.

The Singapore Computer Society, which represents more than 60,000 ICT professionals across the island, has recommended that platform operators adopt layered detection — combining hash-based matching with machine-learning classifiers trained on locally sourced datasets. The society's position, outlined in a technology guidance document published in early 2026, argues that reliance on any single detection method leaves gaps that bad actors quickly learn to exploit.

At Ngee Ann Polytechnic's Centre for Applied Research, practitioners have been running pilot programmes with small and medium-sized e-commerce merchants along Kallang Avenue to test automated image-authenticity checks at the point of upload. The results, presented at a digital trust forum held at the Singapore Management University campus on Stamford Road in March 2026, suggested that real-time flagging reduced duplicate image submissions by roughly 40 percent in participating merchant accounts over a three-month trial period. That figure has not been independently audited but has been cited approvingly by agency officials in subsequent briefings.

Policy Direction and Practical Pressure

The IMDA's Digital Trust Centre, established in 2023 under the Digital Connectivity Blueprint, is the closest thing Singapore has to a central clearing house for standards on this issue. It has been consulting with platform operators — including Carousell, PropertyGuru and government-linked portals — on voluntary baseline standards for image provenance. Industry sources familiar with the consultation, speaking on background because the process is ongoing, say a formal guidance note is expected before the end of the third quarter of 2026.

Carousell, which hosts millions of listings and is headquartered at one-north in Buona Vista, has previously described its automated content moderation systems to trade media without disclosing detection rates. PropertyGuru, listed on the New York Stock Exchange, includes image-quality checks in its agent compliance framework under the Council for Estate Agencies' regulatory oversight.

For ordinary Singaporeans — whether selling a sofa on a classifieds app or buying a resale flat in Punggol — the practical advice emerging from this debate is straightforward: submit original photographs taken on the day of listing, retain metadata where possible, and report suspicious duplicate imagery through platform flagging tools rather than assuming automated systems will catch everything. Agencies and platforms alike are signalling that user-level vigilance remains part of the solution, even as backend detection grows more sophisticated.