Hong Kong Region overview

Benchmark Drilldown · Updated 2026-05-04

Hong Kong Governance Model

Hong Kong's AI governance model can be summarised as: Voluntary guidelines, no dedicated law.

Governance model
Voluntary guidelines, no dedicated law
Strategy year
2022
Region
Hong Kong

Detail

Hong Kong's AI governance is "voluntary guidelines plus parallel agencies": (1) the Digital Policy Office (DPO) issued the Artificial Intelligence Ethical Framework in 2024 — originally for government bureaux, later revised for general organisational use — covering guiding principles, leading practices, and an AI assessment template; (2) DPO released a Generative AI Technical and Application Guideline in April 2025 to address GenAI-specific risks (data leakage, model bias, hallucination); (3) the Privacy Commissioner for Personal Data (PCPD) separately published an AI Model Personal Data Protection Framework in June 2024, anchored in PDPO compliance. Three frameworks coexist but cover separate slices — DPO handles ethics and best practice, PCPD handles privacy compliance — and Hong Kong lacks a unified testable toolkit comparable to Singapore's AI Verify. Assessment: governance is "soft and dispersed", with weaker prescriptive clarity than Singapore, but the response speed (GenAI guideline shipped April 2025) is not slow.

References

Continue exploring this region

Data note

Curated benchmark drilldown profile, last updated 2026-05-04.