On April 15, 2025, the Hong Kong Government's Digital Policy Office released the Hong Kong Generative Artificial Intelligence Technical and Application Guideline, aimed at providing operational guidance for stakeholders involved in generative AI.

The Guideline proposes five key principles of governance, emphasizing compliance with laws, security and transparency, accuracy and reliability, fairness and objectivity, and practicability and efficiency, alongside recommendations for user-friendly fact-checking tools.

A four-tiered risk classification system is introduced, categorizing AI systems as 'unacceptable risk,' 'high risk,' 'limited risk,' or 'low risk,' each accompanied by specific regulatory recommendations.

For systems classified as 'unacceptable risk,' which could endanger human safety, the Guideline suggests full prohibition and legal liability, while 'low risk' systems only require self-certification of risks.

Service users are encouraged to take responsibility for the content generated by AI, which includes disclosing AI involvement and verifying the authenticity of the content.

The Guideline recommends that technology developers establish internal teams dedicated to security and compliance, conduct comprehensive testing, and perform regular compliance reviews.

As AI technology continues to evolve, there may be a need for new statutes or amendments to existing laws to keep pace with ongoing developments.

While generative AI can automatically create content such as text, images, and audio, it also poses unpredictable security risks and ethical challenges, highlighting the necessity for best practices in secure development.

The Guideline outlines a governance framework based on five dimensions: personal data privacy, intellectual property, crime prevention, reliability and trustworthiness, and system security.

The Guideline specifically targets three main stakeholder groups: technology developers, service providers, and service users, clearly outlining their roles and responsibilities in the application of generative AI.

To ensure responsible use of generative AI, service providers are urged to create a framework that prevents the production of illegal content and effectively manages personal data privacy.

Reflecting a non-binding approach, the Guideline relies on existing legal frameworks to address specific AI-related legal issues while promoting a balance between innovation and compliance.