NST Leader: Finally, AI governance

Malaysia's maiden AI governance framework is poised to be introduced to Parliament in the first quarter of 2027. Lawmakers must strike a balance between safeguarding the public and not hampering innovation. This step is timely, yet AI is already integrated into various sectors including business, education, healthcare, finance, government and daily life.

Malaysians are increasingly encountering AI-generated scams, impersonation, deepfakes and misinformation. The Bill should not be perceived as an attempt to slow down AI, but rather a means to foster trust in its implementation. Not every AI application carries the same risk. Lighter regulatory burdens should apply to low-risk uses, while stricter requirements should be imposed on high-risk or high-impact AI, particularly in domains such as healthcare, finance, credit decisions, employment and recruitment, education, policing and public safety, critical infrastructure, and public services.

There are ample reasons to adopt a cautious approach, with AI risk management being the most practical reason. Several countries like the UK, Singapore, Canada, Japan, South Korea, and a few others have started with businesses, education, and commercial arrangements between themselves. Singapore, Malaysia's closest neighbor, did not initiate its AI journey with a single invention; it was a gradual process that began with computerization, university research, government initiatives, and national strategies.

Singapore's interest in computing is not a recent development. In 1963, the Central Provident Fund (Malaysia's EPF) automated its administrative work, leading to increased computerization within the civil service. This was followed by the National Computerisation Plan, Computerisation Masterplan, and research by the country's leading universities.

Now, Singapore has established the National AI Council, positioning artificial intelligence as a cornerstone of its technology strategy. Malaysia must ensure human accountability in AI systems. Machines are meant to assist humans, not serve as an excuse for organizations to evade responsibility. When an AI system makes or significantly influences a decision that affects a person's job, loan, healthcare, or rights, meaningful human oversight is essential.

Someone must ultimately be held accountable when AI causes harm. The law must grant Malaysians the right to know when AI is involved.