About Hong Kong's Regulatory Framework for Ethical and Efficient AI Data Usage

As we navigate the complexities of information usage, it is crucial to ensure that data practices do not infringe on copyrights or engage in illegal activities. This raises an important question: how can we effectively regulate AI tools, particularly generative AI, through legal frameworks in Hong Kong?

According to Lexology.com, the current AI regulatory framework in Hong Kong encompasses several key areas:

1. Data Protection and Privacy: Guidelines on the processing of personal data by AI systems.
2. Intellectual Property (IP): Issues related to IP infringement and ownership of AI-generated content.
3. Operation and Deployment of AI Systems: Current positions and future directions for AI in Hong Kong.
4. Potential Bias in AI Systems: Existing safeguards under Hong Kong laws.
5. Industry Guidance: Sector-specific guidance from the Hong Kong Monetary Authority (HKMA) and the Securities and Futures Commission (SFC). Data Protection and Privacy

The regulation of AI and its ethical usage has been a topic of discussion, particularly since the introduction of the Personal Data (Privacy) Ordinance (PDPO) in Hong Kong. The Office of the Privacy Commissioner for Personal Data (PCPD) has also provided guidance on the ethical development and deployment of AI systems.

Their Data Protection Principles (DPP) include: ? Purpose and manner of use ? Accuracy ? Security ? Openness ? Access and correction

In August 2021, the PCPD outlined three core values for AI usage:

1. Respectful
2. Beneficial
3. Fair

These values align with seven internationally recognized principles, which include accountability, human oversight, transparency, data privacy, fairness, beneficial AI, and reliability.

To implement these values effectively, organizations should focus on:

? Establishing AI strategy and governance

? Conducting risk assessments

? Engaging stakeholders In June 2024, the PCPD released a Model Framework, further detailing recommended measures for data protection and privacy in AI.

However while these guidelines may not be compiled as mandatory, what are possible cocerns?

Intellectual Property Rights

1, Potential Infringement of IP

IP rights in Hong Kong are governed by the Copyright Ordinance, Trademarks Ordinance, and Patents Ordinance. If a developer uses copyrighted materials to train AI systems, it could result in copyright infringement. However, the legal landscape remains ambiguous regarding the use of prompts in generative AI that results in outputs resembling existing trademarks, such as the Coca-Cola logo.

2, Ownership of AI-Generated IP

AI systems now produce a wide array of content, leading to questions about ownership of the resulting IP. In various jurisdictions, including the UK and across Asia, laws stipulate that inventors must be natural persons.

Can the inventor / author be AI?

Korea – Patent Act requires the inventor has to be a person

Japan – Patent Act requires the inventor has to be a natural person with Shimei

Taiwan – Patent guideline requires a natural person, with name and nationality

China – Patent Law & guideline requires the inventor has to be a person

Hong Kong – Patent Ordinance requires a human inventor

Philippines – Patent law requires a natural person inventor

Vietnam – copyright law requires AI user as the inventor nomatter what degree of work

Thailand & Malaysia – Patent law requires the inventor to be individuals

India – Patent law requires the inventor to be individuals

Australia – Full Federal requires it must be a natural person

New Zealand – Patent Acts rules that AI cannot be an inventor

3, Operation and Deployment of AI Systems The European Union's proposed Artificial Intelligence Act adopts a risk-based approach to regulate AI systems, distinguishing between high-risk and lower-risk applications. In contrast, Hong Kong is beginning to implement measures to monitor AI usage while promoting ethical AI guidelines. This balance fosters innovation but also raises concerns about potential risks.

4, Addressing AI Bias AI systems, particularly those using large datasets, may inadvertently perpetuate human biases. Hong Kong's anti-discrimination laws provide some protections, but they may not fully address the nuanced complexities of AI-generated bias. AI algorithms can play a role in identifying and mitigating these biases to promote fairness.

5, Industry Guidance The HKMA's 2019 circular on "High-level Principles on Artificial Intelligence" emphasizes governance, application design, and ongoing monitoring. It encourages authorized institutions to adopt risk-based approaches to minimize risks associated with AI applications. The SFC has also highlighted the importance of responsible technology deployment, emphasizing the necessity of thorough testing and robust governance. Ultimately, accountability lies with humans, not AI. Conclusion In recent discussions, officials have emphasized the need for responsible data collection, IP protection, and bias mitigation in AI operations.

Conclusion: As Hong Kong continues to adapt its regulatory framework, it is essential to remain vigilant and optimistic about the future of AI in the region.

Sources: Key points of AI data protection and privacy regulations in Hong Kong - Leaders in Law An Overview of the AI Regulatory Framework in Hong Kong - Lexology