Malaysia’s National AI Governance and Ethics Guidelines (28.09.24)

Authored by: Mr Archak Das

On 20^th September, 2024, Malaysia’s Ministry of Science, Technology, and Innovation (MOSTI) took a significant step forward in shaping the future of artificial intelligence (AI) within the country by publishing the National Artificial Intelligence Governance and Ethics Guidelines (AIGE). As AI continues to play a transformative role across industries, governments worldwide are recognizing the need for responsible governance and Malaysia is no exception, where the AIGE serves to provide a comprehensive framework for the safe and ethical deployment of AI technologies. The AIGE guidelines are designed to cater to the needs of three key stakeholder groups, i.e. end users of AI, policymakers in government agencies and organizations, and developers and technology providers. Through this approach, the guidelines aim to ensure that AI technologies are used responsibly, mitigate associated risks, and promote transparency and fairness.

I. Key Considerations for National AI Guidelines Development

Inclusivity and Accessibility: The National Guidelines aim to encourage an inclusive AI ecosystem, ensuring that all stakeholders, including the public, are represented and no one is left behind in the AI landscape.
Building Trust: These Guidelines serve as a commitment by the government to regulate AI responsibly, encouraging trust in the nation by providing clear governance measures.
Clarity and Comprehensibility: Designed to be easily understood, the Guidelines ensure that all stakeholders can comprehend and engage with AI policies.
Global and Local Alignment: The Guidelines align with national, regional, and global AI practices, while integrating local and indigenous knowledge, ensuring consistency with worldwide standards.
Synergy Across Sectors: Collaboration between government agencies and stakeholders is emphasized to operationalize AI effectively within national policies.
Call to Action for Sectors: The Guidelines serve as a framework, urging sectors like healthcare and finance to develop specific AI implementation strategies.
Living Document: Recognizing AI’s dynamic nature, the Guidelines will undergo periodic review and updates to stay relevant to evolving technologies and stakeholder needs. They will be made available in both Malay and English.

II. Objectives of National Guidelines AIGE

The objectives of the National Guidelines for AI Governance and Ethics (AIGE) in Malaysia, the document is crafted to align with the country’s National AI Roadmap 2021-2025 and ensure the responsible development of AI. The guidelines aim to encourage trust in AI systems by adhering to seven core AI principles and managing risks associated with AI deployment. While voluntary, the guidelines encourage the broad adoption of responsible AI across various stakeholders, including developers, designers, end users, policymakers, and technology providers, with the ultimate goal of enhancing national productivity, economic growth, and competitiveness. The overarching theme—AI for Malaysia, AI for All—reflects the guidelines’ inclusive vision for AI’s role in advancing the nation’s future.

III. Seven Core AI Principles

In the National Guidelines for AI Governance and Ethics (AIGE), seven core AI principles are established to promote the responsible development and deployment of artificial intelligence technology. These principles serve as a framework to ensure AI systems are developed ethically, safely, and with a human-centered focus.

Fairness: AI must be developed to avoid bias or discrimination, ensuring equitable access and opportunities for all, irrespective of age, gender, religion, or ethnicity. AI solutions should not be one-size-fits-all and should be inclusive of diverse user needs.
Reliability, Safety, and Control: AI systems must function as intended and be resilient against unauthorized access. Ensuring privacy, data protection, and system integrity throughout the AI lifecycle is critical. Developers should implement international standards for security and privacy, while users’ consent should be respected regarding data use.
Privacy and Security: The systems must be strong and thoroughly tested for safety and reliability. This principle emphasizes the need for AI systems to handle unexpected scenarios while maintaining control. Proper testing, certification, and risk mitigation efforts are encouraged, especially in high-risk sectors such as autonomous vehicles and military applications.
Inclusiveness: AI must be developed with an inclusive approach that addresses the needs of all societal segments, particularly vulnerable groups. AI solutions should be designed to benefit everyone, considering national needs and experiences, and developers should embrace diversity and inclusivity in AI decision-making processes.
Transparency: Transparency is crucial to AI governance, ensuring AI algorithms and their decision-making processes are open for review and explanation. Stakeholders must be able to understand how AI systems work, the purpose of their use, the data they were trained on, and how decisions can be challenged.
Accountability: Developers, AI owners, and operators must be accountable for their AI solutions. This includes taking responsibility for both successes and failures, and ensuring that the systems comply with ethical standards and governance. Key elements such as system purpose, technological capability, quality, and impact on sensitive users must be addressed.
Pursuit of Human Benefit and Happiness: AI systems should prioritize enhancing human welfare and happiness, aligning with human-centered values. AI should be designed for the betterment of society, with mechanisms such as human-in-the-loop (HITL), human-on-the-loop (HOTL), and human-in-command (HIC) to ensure human oversight and prevent negative outcomes.

These principles collectively emphasize building trustworthy AI systems that serve humanity’s best interests, ensuring fairness, accountability, and transparency while promoting safety and inclusion.

IV. Human Friendly AI Platform

The Human-Friendly AI (HF-AI) Platform is an essential component in Malaysia’s effort to enhance AI development and adoption. Designed for operational efficiency and seamless user interaction, it provides immediate, 24/7 support to address queries, solve problems, and offer information without any delays or downtime. The platform enables automation of routine tasks, allowing human employees to focus on more complex strategic activities, which improves productivity while reducing operational costs.

V. Guidelines for the Three Stakeholders-

End users: One of the main objectives of the AIGE is to ensure that AI users understand their rights and responsibilities when engaging with AI technologies, as with this rapid advancement of AI applications, from virtual assistants to personalized marketing, users often interact with AI systems without fully comprehending the implications. The guidelines emphasize the importance of consumer protection principles that govern the interaction between users and AI technologies and a core focus of the guidelines for users is data privacy and protection. AIGE recommends that users should take proactive steps to safeguard their personal information themselves as well. The guidelines also provide recommendations on best practices when adopting AI technologies, which includes understanding the limitations of AI systems and how to identify biases in AI outputs, ensuring that users are aware of potential risks of using AI technology.
Policymakers of Government Agencies, Organizations, and Institutions: Policymakers play a crucial role in shaping the landscape of AI governance and according to the AIGE, they are tasked with the responsibility of encouraging technological innovation while ensuring accountability and fairness. The recommendation provide for them is two fold that is first to embed a human-centric approach in the development and use of AI. This helps to ensure that AI technologies are designed to serve human needs while respecting human rights, focusing on transparency and fairness. Second, to develop comprehensive AI governance frameworks that address various dimensions, including risk mitigation.

Developers, Technology Providers, and Suppliers: Developers, technology providers, and suppliers are at the forefront of AI innovation as they bear a significant responsibility to ensure that the AI systems they design, build, and deploy are fair, accountable, and transparent. One of the primary concerns is the issue of algorithmic bias as it can lead to unfair or discriminatory outcomes, affecting marginalized communities or reinforcing existing societal inequalities. The guideline calls for developers to mitigate biases during the design and development phases and this involves testing AI models against diverse data sets and adopting techniques to ensure that outputs are representative and fair. The next recommendation for them is to monitor the AI performance, ensuring that the AI systems function as intended and comply with ethical standards. Lastly, data sharing between developers. The access to larger and more diverse data sets allows for the creation of AI systems that are more accurate and less prone to bias, but the AIGE also emphasizes that data sharing must be done in a manner that respects data privacy and security regulations, ensuring that the rights of individuals are protected.

VI. The Importance of AI Governance in Malaysia’s Technological Landscape

The publication of the National Artificial Intelligence Governance and Ethics Guidelines by MOSTI reflects Malaysia’s commitment to encourageing a responsible AI ecosystem and as AI continues to grow into various sectors, from healthcare and education to finance and transportation, the need for a comprehensive governance framework becomes increasingly critical. Keeping the focus on the rights and responsibilities of users, encouraging ethical development practices, and guiding policymakers in creating balanced regulatory environments, the AIGE serves as a blueprint for AI governance in Malaysia and it is a significant milestone in ensuring that the rapid growth of AI technology is matched with responsible and ethical practices.

Conclusion

As AI becomes an integral part of Malaysia’s digital future, the National Artificial Intelligence Governance and Ethics Guidelines published by MOSTI will play a pivotal role in shaping the country’s AI landscape as these guidelines provide a clear framework for end users, policymakers, and developers to ensure the safe and ethical use of AI technologies. Whether it’s the emphasis on data privacy, the focus on human-centric AI development, or the push for responsible innovation, the AIGE ensures that AI growth in Malaysia remains aligned with the broader goals of transparency, fairness, and accountability.