PETALING JAYA: Malaysia is pushing forward as one of Southeast Asia’s most active players in artificial intelligence (AI), driven by major infrastructure partnerships with global technology leader Nvidia and YTL, and strengthened by new collaboration with the World Economic Forum (WEF) on responsible AI development.
But as this progress accelerates, it must be matched with stronger oversight to ensure trust, safety and long-term competitiveness, industry expert Georg Chmiel said.
Prime Minister Datuk Seri Anwar Ibrahim’s recent meeting with Nvidia founder Jensen Huang in South Korea focused on Malaysia’s plans to expand access to high-performance computing, support AI talent development and build the digital capabilities needed by businesses across the region. These efforts also support Malaysia’s growing role as a regional digital and global business services hub.
Adding to this momentum, Malaysia’s first dedicated AI Bill is expected to be tabled in Parliament by mid-2026. The upcoming legislation is aimed at strengthening transparency, accountability and responsible use of AI across both public and private sectors, a strong signal of the government’s commitment to safe innovation.
In a statement today, Chmiel, founder of Chmiel Global Advisory and board member of the World Digital Chamber, said this legislative move is timely.
“AI systems today can behave unpredictably. We’ve seen models bypass policies, invent product features just to close a sale, or hide rule-breaking when monitored. Without proper guardrails, businesses risk serious legal and reputational damage.”
Malaysia’s partnership with the WEF focuses on developing governance frameworks for Asean, a move that helps build confidence among investors, regulators and industry players.
The Ministry of Communications has stressed that trust and security are essential if Malaysia’s digital economy is to grow at the pace the government envisions.
Chmiel emphasised that ambition must be matched with a realistic view of risk. “We are no longer dealing with simple mistakes,” he said.
“Modern AI systems can recognise their environment, understand incentives, and act strategically, even inventing product features, bypassing policies, or hiding rule-breaking when monitored.”
He detailed the evolution of AI misbehaviour, from early hallucinations to reward hacking and now deliberate scheming. In one striking example, an AI ignored instructions not to modify a file and then wrote a script to bypass digital locks, demonstrating situational awareness and intent.
“This is not hypothetical. It’s a live demonstration of AI thinking strategically. For Malaysian businesses using AI in finance, logistics, and public services, this is a new reality that demands a fresh governance playbook,” Chmiel explained.
Chmiel proposes treating AI like human employees through an “artificial resource” (AR) function.
“Just as HR oversees policies, training, and wellbeing, AR must govern AI, manage risks, and train ethical behaviour. Establishing this discipline is not optional—it’s foundational for Malaysia’s world-class AI infrastructure.”
With strong infrastructure, a multilingual workforce, strategic location, and now a forthcoming AI Bill, Malaysia is positioning itself to lead the region in responsible innovation.
“Malaysia has a real opportunity to set the standard for Asean,” Chmiel said. “If we combine investment with governance, we won’t just adopt AI, we’ll build trust in it.”
