Hong Kong's War on Fake Images: What Officials, Experts and Industry Figures Are Saying

Hong Kong's digital regulators and technology industry insiders are sharpening their positions on duplicate image manipulation, with calls intensifying this summer for a unified framework to govern how copied, cloned or AI-generated visuals are identified, flagged and removed across local platforms and public-sector systems.

The push comes as AI image-generation tools have become widely available and cheap, flooding everything from property listings on Midlevels agency websites to public health announcements distributed via government portals. The Office of the Communications Authority has fielded an increasing volume of complaints in recent months relating to misleading digital content, though the authority has not published a specific breakdown of image-related cases to date.

Why the Debate Has Sharpened Now

The immediate trigger is a combination of factors converging in mid-2026. The government's Smart City Blueprint 2.0, which covers digital infrastructure priorities through 2027, includes provisions on data integrity that officials at the Innovation, Technology and Industry Bureau have publicly described as extending to visual content standards — though no specific legislative amendment has been tabled in the Legislative Council as of this week.

Hong Kong Cyberport, the tech hub in Pok Fu Lam that houses more than 2,000 companies, has been running a series of workshops since March on responsible AI deployment. Practitioners there have flagged that duplicate image detection — identifying when the same image is reused with minor alterations to evade automated filters — remains a gap in current content moderation pipelines used by local e-commerce and fintech companies.

The Hong Kong Applied Science and Technology Research Institute, known as ASTRI and headquartered in Science Park in Pak Shek Kok, has been developing hash-based image fingerprinting tools that can match near-duplicate visuals even after cropping or colour adjustment. ASTRI has described the technology as ready for pilot deployment, though no public contract with a government department has been announced.

The concern is not abstract. In the financial services sector, the Securities and Futures Commission issued guidance in January 2026 noting that misleading marketing materials — including images used in investment product promotions — fall within its existing regulatory perimeter. The SFC noted at the time that digital assets and tokenised product marketing warranted particular scrutiny, given the speed at which promotional content circulates on platforms including Telegram channels popular among retail investors in Wan Chai and Mong Kok.

Where Industry and Government Diverge

Not everyone agrees on method. Some technology lawyers practising in Central have argued that imposing mandatory duplicate-detection requirements on platforms would create compliance costs disproportionate to the harm, particularly for smaller operators. They point to the existing Personal Data (Privacy) Ordinance framework as already offering some remedies where fake images involve identifiable individuals.

Consumer advocacy groups, including the Consumer Council, have taken a harder line, pressing for visible labelling on AI-generated or recycled images used in commercial contexts — a standard that has gained traction in the European Union but has no direct equivalent in Hong Kong law as of July 2026.

The Hong Kong Productivity Council, which advises small and medium enterprises across districts including Kwun Tong and Tsuen Wan, has circulated guidance on image metadata verification practices. The council has recommended that businesses adopt perceptual hashing tools as a baseline, citing estimates from international cybersecurity research suggesting that a significant share of online product images in Asia-Pacific markets involve some degree of duplication or manipulation — though the council has not attributed a specific local figure.

What comes next will likely hinge on whether the Innovation, Technology and Industry Bureau moves duplicate image standards into its forthcoming review of the Code on Access to Information, expected in the fourth quarter of 2026. Industry bodies including the Hong Kong Internet Registration Corporation have already written to the bureau requesting a formal consultation. Until then, businesses would be wise to audit their own content pipelines now, given the direction regulators appear to be travelling — and the reputational exposure that comes with being caught on the wrong side of the line.