[IP Alert] How the Malaysian National Guidelines on AI Governance and Ethics Will Impact You
InsiderTAPS October 2024

How the Malaysian National Guidelines on AI Governance and Ethics Will Impact You
INTRODUCTION
While generative artificial intelligence (“GenAI”) has existed for some time, the year 2023 marked the beginning of a substantial increase in the use of GenAI when the world witnessed the capability of ChatGPT and other GenAI models such as Copilot and Bard (renamed Gemini). One may say that these models have been transformative and have brought drastic changes to many aspects of our work and personal lives.
To encourage and enhance ethical and responsible AI development and deployment, the Malaysian government through its Ministry of Science, Technology, and Innovation (“MOSTI”) launched the National Guidelines on AI Governance & Ethics (“AIGE”) on 20 September 2024. [1] AIGE is intended to support the implementation of the National Artificial Intelligence Roadmap 2021-2025 (“AI-Rmap”) published by MOSTI in August 2022.
This article aims to highlight some of the salient aspects of the AIGE and compare them with the related laws and regulations in a few other jurisdictions.
AIGE: SELF-REGULATION GUIDELINES OR LEGAL OBLIGATIONS?
The AIGE are voluntary guidelines for stakeholders which aim to, among others, educate stakeholders about and encourage them to comply with the Seven Key Principles of AI. The stakeholders are broadly categorised as the end users, policymakers and AI technology providers.
Notwithstanding the voluntary nature of the AIGE, it is pertinent to note that the Minister of MOSTI, YB Tuan Chang Lih Kang, said during the launch of the AIGE that while the AIGE have no legal effect and only “act as a reference for industries and AI players” the AIGE should be viewed as “as a discipline to adhere to” and he further expressed hope that certain aspects of the AIGE will eventually be legislated. [2] The aforesaid aligns precisely with the government’s recent introduction of and amendments to legislation relating to the technology sphere, such as the new Cyber Security Act 2024 and the Personal Data Protection (Amendment) Bill 2024 which seeks to amend the Personal Data Protection Act 2010. Although MOSTI has not proposed a legislation specifically catered to AI, certainly, the current and future versions of AIGE will likely gradually shape the future AI laws and regulations in Malaysia.
SEVEN KEY PRINCIPLES OF AIGE
The AIGE are drafted based on the following Seven Key Principles of AI. It is suggested that these principles will be important references for future policymaking and discussions related to AI.
Principle | Elaboration |
Fairness | AI development should avoid bias and discrimination including unintentional biases in data processing. The benefits of AI should be equally distributed for all races, genders and religions without leaving anyone behind. |
Reliability, Safety and Control | AI systems should be safe and secure, and perform as intended, and resist being compromised by unauthorised parties. |
Privacy and Security | AI systems should be safe and encourage trust in humans with automatic measures to prevent data privacy breaches such as rigorous testing, monitoring, and implementing fail-safe mechanisms. Autonomous safeguards such as obtaining informed consent, ensuring data security and defining the permissible uses of personal information should be mandated and prepared by AI systems developers and providers. This is to minimise any potential harm, especially in high-risk areas concerning military and healthcare applications. |
Inclusiveness | AI systems should be inclusive to provide equal access to AI that will address national needs and mitigate social gaps. For example, AI search engines should include user feedback to improve search and chat results that are automatically embedded in applications. |
Transparency | Transparency principles are essential in the decision-making process with five elements to be adhered to: (a) full disclosure about the facts that an AI system is being used in decision-making (b) the intended purpose of the AI systems (c) the data used for training and potential biases exist in the data (d) maintenance and assessment of AI systems (e) the ability to challenge the decisions made by AI systems |
Accountability | This principle requires that AI developers and providers are accountable for the issues and implications of the technology. Policymakers should design a legal framework which assigns responsibility and mechanisms for AI systems to different stakeholders. |
Pursuit of Human Benefit and Happiness | The main objective of AI technology is to enhance societal values for the betterment of individuals and communities by prioritising human-centred values. AI systems creators should design human oversight mechanisms, such as: (a) Human-in-the-loop (HITL) which refers to the capability for human intervention in every decision cycle of the AI system; (b) Human-on-the-loop (HOTL) which refers to the capability for human intervention during the design cycle of the system and monitoring the system’s operation; and (c) Human-in-command (HIC) which refers to the capability to oversee the overall activity of the AI systems and the ability to decide when and how to use the AI system. |
AI GOVERNANCE IN THE EUROPEAN UNION (“EU”) AND CHINA
As we observe the development of future AIGE in Malaysia, it may be of interest for stakeholders to refer to two other jurisdictions, such as the European Union (“EU”) and China, to prepare themselves for the possible directions that Malaysia might take for AI governance and ethics.
The EU launched its legislation for AI on 12 July 2024, namely the European Union’s Artificial Intelligence Act (“EU AI Act”). [3] The EU AI Act came into force on 1 August 2024, but its effective date varies for different segments over a period of 24 months. With the extra-territorial nature of the EU AI Act as stated in Article 2(1), [4] Malaysian entities may be affected; for instance, if they (with or without a physical presence in the EU) make the output of an AI system available for use in the EU. The EU AI Act does not apply to, among others, AI systems utilised for military, defence or national security [5] and individuals using AI systems for personal use. [6]
The EU AI Act is a specific legislation that governs the development, deployment, supply, and use of AI systems based on the risk level of the systems. Malaysia’s AIGE are a preliminary introduction to AI governance with self-regulatory obligations for end users, policymakers and developers of AI systems. However, even the EU AI Act does not deviate from the Seven Key Principles of AI. For instance, under Chapter IV, Article 50 of the EU AI Act, providers and deployers must disclose to users when they are interacting with or consuming content generated by an AI system. [7] This aligns with the transparency principle propounded by the AIGE.
Shifting our lens now to the other side of the globe, China. China is probably one of the more progressive jurisdictions when it comes to regulating new technologies and that is demonstrated by the launch of its Interim Measures for the Management of Generative Artificial Intelligence Services on 13 July 2023 (“AI Measures”) for GenAI soon after the world saw the capabilities and vast potential of ChatGPT. The AI Measures came into force on 15 August 2023. [8] Similar to the EU AI Act, the AI Measures extend beyond domestic territory to include foreign entities including deployers which use GenAI system to provide services to the public in China.[9] Consisting of 24 provisions, the AI Measures introduce the dos and don’ts of GenAI-related activities. With resemblance to the AIGE, the AI Measures require that all data processing and training are done lawfully with consent from intellectual property rights and personal information holders. Article 4 expressly requires the provision and use of GenAI services to practise fairness, respect intellectual property rights and business ethics, and enhance the transparency, accuracy and reliability of the content provided by the GenAI services. [10] Additionally, the AI Measures mandate providers to promptly remove any illegal content and adopt AI optimization measures. [11]Providers must establish intuitive reporting mechanisms to the relevant authorities. [12]
In short, it appears that most AI-related regulations and guidelines are aligned across different jurisdictions with the Seven Key Principles of AI as the foundation.
CONCLUSION
Besides the AI-Rmap, the AIGE are an important milestone for Malaysia in framing the big picture of AI governance and ethics. It is hoped that legislation related to AI and intellectual property will soon be enacted or amended, as the case may be, to ensure that issues and gaps in the existing legislation arising from technological advancement will be adequately addressed. To achieve this, it is crucial for the relevant ministries and agencies of the government to engage in more collaborative efforts with all stakeholders, including AI system developers and providers, legal fraternity (both private and public) and academics.
Although the AIGE are not legally binding, it may well be prudent for businesses involved in this space, especially those with aspirations to have cross-jurisdictional operations beyond Malaysia, to closely observe the requirements of AIGE, since doing so would be a big step towards future regulatory compliance.
[1] Ministry of Science, Technology, and Innovation (MOSTI), ‘AI Governance and Ethics Guidelines (AIGE)’ <https://www.mosti.gov.my/dasar/#flipbook-df_74045/1/>
[2] The Star, ‘AIGE to serve as initial reference for industries, AI players, not legal requirement, says Tech Minister’, 20 September 2024 <https://www.thestar.com.my/news/nation/2024/09/20/aige-to-serve-as-initial-reference-for-industries-ai-players-not-legal-requirement-says-tech-minister>
[3] Artificial Intelligence Act (Regulation (EU) 2024/1689) (European Union).
[4] Ibid, Article 2(1).
[5] Ibid, Article 2(3).
[6] Ibid, Article 2(10).
[7] Ibid, Article 50.
[8] Ibid, Article 24.
[9] Ibid, Articles 2 and 20.
[10] Ibid, Article 4(5).
[11] Ibid, Article 14.
[12] Ibid, Article 15.
Written by:
Ng Kim Poh
Partner
T: +603 2050 1870
Rachel Tan
Pupil