KUALA LUMPUR: Khazanah Research Institute (KRI) has proposed setting up a central national artificial intelligence (AI) agency to oversee and implement Malaysia's AI strategy, along with creating an agile regulatory framework capable of adapting to the country's AI adoption.

In a statement today, KRI said these recommendations were unveiled during its inaugural Artificial Intelligence Impact and Governance (AIIG) Roundtable, held on Aug 27, which brought together key stakeholders, including policymakers, public officials, private sector representatives, academics, and civil society organisations, to discuss the future of AI governance in Malaysia.

"AI governance stakeholders agreed on the need for national coordination, with a clear and unified strategy that leads to systematic guidelines or regulations and straightforward implementation processes.

"KRI proposes that the government establish a central national AI agency to serve as the single focal point for coordinating and implementing Malaysia's strategy on AI adoption and governance, while also building connections between experts across different sectors and gathering and disseminating information about AI impacts to the public," the statement said.

KRI further said Malaysia can learn from other countries that are already developing AI guidelines and regulations by building international learning partnerships and participating in international standards-setting processes.

"However, global regulations should not be adopted wholesale into a Malaysian context, and KRI recommends ensuring an agile regulatory framework within Malaysia's current context as a deployer of AI, capable of adapting in tandem with Malaysia's AI adoption, while also setting clear national objectives for this framework to ensure policy coherence across and within sectors.

Meanwhile, KRI said its research highlighted three significant risks associated with AI, which are falling behind globally due to non-adoption, experiencing unintended consequences from unsafe AI, and the weaponization or misuse of AI.

"Thus, KRI recommends upskilling policymakers and the civil service on both how to use and govern AI," KRI said.