HKMA Steps Towards Regulating Big Data Analytics, AI in FinTech

Herbert Smith Freehills says the HKMA push for adoption of the Ethical Accountability Framework in FinTech development represents a broader movement toward regulating artificial intelligence.

The HKMA (Hong Kong Monetary Authority) recently issued a circular to encourage authorised institutions to adopt the “Ethical Accountability Framework” (EAF) for the collection and use of personal data issued by the Office of the PCPD (Privacy Commissioner for Personal Data).

A report on the EAF was published by the PCPD in October 2018, which explored ethical and fair processing of data through (i) fostering a culture of ethical data governance and (ii) addressing the personal data privacy risks brought by emerging information and communication technologies such as big data analytics, artificial intelligence and machine learning.

The EAF is expressly stated to be non-binding guidance, intended as a first step towards a privacy regime better equipped to address modern challenges. However, the HKMA’s circular arguably elevates the legal status of the EAF for authorised institutions.

The HKMA is likely to incorporate the EAF into its broader supervision and inspection of authorised institutions. In particular, in construing the principles based elements of the Supervisory Policy Manual as it applies to FinTech, the EAF will undoubtedly have an influence going forward.

Tension between the value of data-processing technology and public trust

Big data has no inherent value in its raw form. Its value lies in the ability to convert that data into useful information for organisations, which can then generate knowledge or insight relating to clients or the market as a whole through data analytics or artificial intelligence.

Ultimately, this insight results in competitive advantage. However, a tension exists between (i) developing data-processing technology to gain a competitive advantage; and (ii) addressing public distrust arising from the data-intensive nature of such technology.

As the EAF report highlights, the existing regulatory regime in Hong Kong does not adequately address the privacy and data protection risks that arise from advanced data processing. Big data analytics and artificial intelligence in particular pose challenges to the existing notification and consent based privacy legal framework.

These challenges are not limited to the legal framework in Hong Kong. The privacy and data protection legislations on an international level are also ill-equipped to anticipate advances in data-intensive technology.

Data stewardship accountability

The PCPD sees the need to provide guidance on how institutions could act ethically in relation to advanced data-processing to foster public trust. It reminds institutions to be effective data stewards, not merely data custodians. Data stewards take into account the interests of all parties and consider whether the outcomes of their advanced data processing are not just legal, but also fair and just.

The PCPD also encourages data stewardship accountability, which calls for institutions to define and translate stewardship values into organisational policies, using an “ethics by design” approach. This approach requires institutions to have data protection in mind at every step and to apply the principles of privacy by default and privacy by design.

Privacy by default means that once a product or service has been released to the public, the strictest privacy settings should apply by default. Privacy by design, on the other hand, requires organisations to ensure privacy is built into a system during the entire life cycle of the system.

Ultimately, data stewardship should be driven by policies, culture and conduct on an organisational level, instead of technological controls.

Both the privacy by design and the privacy by default principles are mandatory requirements under the EU’s GDPR (General Data Protection Regulation). The legal development trend is for Asian-based privacy regulators to, whether by means of enacting new laws (e.g. India) or issuing non-mandatory best practice guidance to encourage data users to meet the higher standards under GDPR.

Data stewardship values

The PCPD encourages institutions to adopt the three “Hong Kong Values”, whilst providing the option to modify each value to better reflect their respective cultures. The three Hong Kong Values listed below are in line with the various Data Protection Principles of the Personal Data (Privacy) Ordinance (Cap. 486):

(i) The “Respectful” value requires institutions to:
• be accountable for conducting advanced data processing activities;
• take into consideration all parties that have interests in the data;
• consider the expectations of individuals that are impacted by the data use;
• make decisions in a reasonable and transparent manner; and
• allow individuals to make inquiries, obtain explanations and appeal decisions in relation to the advanced data processing activities.

(ii) The “Beneficial” value specifies that:
• where advanced data-processing activities have a potential impact on individuals, organisations should define the benefits, identify and assess the level of potential risks;
• where the activities do not have a potential impact on individuals, organisations should identify the risks and assess the materiality of such risks;
• once the organisation has identified all potential risks, it should implement appropriate ways to mitigate such risks.

(iii) The “Fair” value specifies that organisations should:
• avoid actions that are inappropriate, offensive or might constitute unfair treatment or illegal discrimination;
• regularly review and evaluate algorithms and models used in decision-making for any bias and illegal discrimination;
• minimise any data-intensive activities; and
• ensure that the advanced data-processing activities are consistent with the ethical values of the organisation.

The PCPD also encourages institutions to conduct “Ethical Data Impact Assessments” (EDIAs), allowing them to consider the rights and interests of all parties impacted by the collection, use and disclosure of data. A process oversight model should be in place to ensure the effectiveness of the EDIA. While this oversight could be performed by internal audit, it could also be accomplished by way of an assessment conducted externally.

International direction of travel

The approach outlined above is not unique to Hong Kong. In fact, at the time the EAF was announced by the PCPD in October 2018, the 40th International Conference of Data Protection and Privacy Commissioners released a Declaration on Ethics and Protection in Artificial Intelligence (Declaration) which proposes a high level framework for the regulation of artificial intelligence, privacy and data protection.

The Declaration endorsed six guiding principles as “core values” to preserve human rights in the development of artificial intelligence and called for common governance principles on artificial intelligence to be established at an international level.

It is clear that there is a global trend toward ethical and fair processing of data in the application of advanced data analytics. For instance, the MAS (Monetary Authority of Singapore) has formulated similar ethical principles in the use of artificial intelligence and data analytics in the financial sector, announced in November 2018.

Another example is the EU’s GDPR’s specific safeguards related to the automated processing of personal data that has, or is likely to have, a significant impact on the data subject, to which the data subject has a right to object. Specifically, a data protection impact assessment assessing the impact of the envisaged processing operations must be carried out before such processing is adopted, if such processing uses new technologies and is likely to result in a high risk to the rights and freedoms of natural persons after taking into account the nature, scope, context and purposes of the processing.

Although this may appear to be a relatively minor development in Hong Kong, we see this as a step in a broader movement toward the regulation of AI and a sea change in the approach to data protection and privacy. The HKMA circular and the EAF are in line with the global data protection law developments, which are largely being led by the EU.

This article was first published by Hannah Cassidy, Jeremy Birch, Sheena Loi and Peggy Chow at international law firm Herbert Smith Freehills.

To Top