Hong Kong Officials and Digital Experts Call for Tighter Controls on AI-Generated Duplicate Images Flooding Online Platforms

Hong Kong's digital governance community is raising urgent concerns about the rapid spread of duplicate and AI-generated images across local news sites, e-commerce platforms and social media — a problem that experts say is outpacing the city's existing regulatory tools. The pressure is mounting on the Communications Authority and the Office of the Government Chief Information Officer to clarify enforcement pathways before the issue compounds ongoing debates about misinformation and intellectual property.

The timing is not arbitrary. Globally, the volume of synthetically generated or algorithmically duplicated images has accelerated sharply through 2025 and into 2026, driven by the widespread adoption of generative AI tools that are freely available to consumers. In Hong Kong, the concern carries extra weight: the city positions itself as a regional data and financial hub competing directly with Singapore, and reputational damage from deepfake or duplicate-image incidents — particularly in financial communications or court filings — could erode confidence in ways that matter to institutional clients across the Greater Bay Area corridor.

What Officials and Industry Figures Are Saying

Advisers linked to the Innovation, Technology and Industry Bureau have privately signalled that revised guidelines on synthetic media are being considered as part of a broader digital economy policy refresh, though no formal consultation document has been published as of July 4, 2026. The Hong Kong Internet Registration Corporation Limited, which manages domain infrastructure and has previously engaged with content integrity questions, is understood to be monitoring international standards bodies for applicable frameworks.

At Cyberport in Pok Fu Lam, several resident companies working in computer vision and media authentication have begun marketing watermarking and provenance-verification tools to local broadcasters and publishers. Professionals at the Hong Kong Applied Science and Technology Research Institute, based in Pak Shek Kok in the New Territories, have flagged the challenge in publicly available technical briefs: conventional hash-based duplicate detection breaks down when AI tools introduce minor pixel-level variations that defeat simple matching algorithms while leaving the image substantively identical to a source original.

Within the legal profession, the Law Society of Hong Kong has noted — in written materials circulated to members, not in direct statements to this publication — that evidentiary rules governing digital images in Hong Kong courts were last substantively reviewed before generative AI tools became mainstream. The practical gap is real: a litigant seeking to challenge an image's authenticity currently relies on expert witness testimony rather than any statutory authentication standard, a situation practitioners describe as expensive and inconsistent.

Scale of the Problem and Pending Policy Steps

Industry research published in Singapore in early 2026 estimated that roughly 38 percent of product images on major Asia-Pacific e-commerce platforms contained some degree of AI manipulation or duplication — a figure that digital rights advocates in Wan Chai have cited when pressing local regulators to engage. Hong Kong Customs and Excise, which handles intellectual property enforcement, has jurisdiction over commercial image theft but currently lacks dedicated technical capacity specifically for AI-duplicate detection, according to public budget documents from the 2025-26 financial year.

The Hong Kong Productivity Council, headquartered in Kowloon Tong, runs a digital transformation support programme for small and medium enterprises that as of mid-2026 does not include modules specifically addressing synthetic image compliance — a gap that business associations in Kwun Tong's industrial belt have flagged to the Trade and Industry Department.

Policy advocates say the most immediately actionable step would be for the Communications Authority to issue a defined category for AI-generated or algorithmically duplicated images within its existing code of practice for television and radio, then extend equivalent guidance to online platforms under the Broadcasting Ordinance review scheduled for the second half of 2026. Absent that, companies operating across the Greater Bay Area face the awkward reality of complying with mainland content-authenticity rules that are stricter in some respects than Hong Kong's own — an inversion that officials on both sides of the boundary have reason to resolve sooner rather than later.