Artificial intelligence has emerged as a powerful tool for supporting decision-making processes, but it cannot and should not displace the fundamental role of human judgment, according to Tun Zaki Azmi, former Chief Justice of Malaysia and current chairman of Pertama Digital Berhad. Speaking at the Wisdom Malaysia 2026 Conference in Kuala Lumpur on August 5, Tun Zaki articulated a nuanced position on the integration of AI within institutional frameworks, particularly in the judiciary where the stakes of decision-making carry profound consequences for justice and the rule of law.
The distinction between AI as an assistive technology and as a decision-making authority carries critical implications for the Malaysian legal system. Tun Zaki emphasised that while artificial intelligence can offer fresh analytical perspectives and streamline preliminary investigative work, the ultimate responsibility for consequential determinations must remain anchored firmly in human hands. This stance reflects an understanding that human judgment encompasses dimensions that technology, no matter how sophisticated, cannot fully replicate—experience accumulated through years of professional practice, ethical reasoning grounded in moral philosophy, and the contextual wisdom that comes from navigating complex situations.
A significant concern Tun Zaki identified relates to AI's inherent technical limitations, particularly the phenomenon of "hallucinations," where artificial intelligence systems generate plausible-sounding but factually incorrect information. In the judicial context, this vulnerability poses severe risks. Judges who rely excessively on AI-generated summaries or analyses without independent verification could inadvertently base their decisions on false premises. This underscores why the traditional judicial approach of careful reading, close interpretation, and personal engagement with statutory text and case law remains indispensable. The personal intellectual labour of a judge cannot be outsourced to algorithms without fundamentally compromising the integrity of the adjudicative process.
Tun Zaki drew upon his extensive experience as former Chief Justice of the Dubai International Financial Centre Courts to illustrate best practices in managing AI's integration into legal work. The DIFC Courts implemented a directive requiring legal practitioners to explicitly disclose whenever artificial intelligence has been deployed in the preparation of legal research, memoranda, or court submissions. This transparency mechanism serves a crucial function: it alerts the presiding judge to exercise heightened scrutiny over materials that may bear the influence of AI processing, allowing the court to assess their reliability with appropriate caution. Such disclosure requirements represent a sensible middle ground that permits beneficial use of AI while maintaining oversight and accountability.
The structural design of appellate courts in Malaysia and elsewhere demonstrates why human plurality remains essential to sound judgment. Tun Zaki noted that appellate panels deliberately comprise an odd number of judges—typically three, five, seven, or nine—specifically to harness the diversity of reasoning and perspective that emerges when multiple experienced minds examine the same question. Judges may genuinely disagree about how statutory language applies to particular facts, or how credibility assessments should weigh in evidentiary evaluations. These divergences reflect the genuine complexity of legal interpretation and factual assessment, not deficiencies. The deliberative process through which judges reconcile their differing views produces jurisprudence that is more robust precisely because it has withstood internal challenge and scrutiny.
This pluralistic approach to judicial reasoning represents an institutional acknowledgment that no single analysis, whether human or artificial, exhausts the possible valid interpretations of law. By designing courts to require consensus among multiple judges, the system builds in resistance to individual error or bias. Artificial intelligence, operating as a deterministic system incapable of true deliberation or principled disagreement, cannot replicate this deliberative function. While AI might identify relevant precedents or summarise evidentiary materials with greater speed than humans, it cannot participate in the reasoned exchange through which judges refine their understanding and arrive at collectively endorsed decisions.
The Wisdom Malaysia 2026 Conference itself reflected broader institutional concerns about knowledge transfer across generational lines. Organised by KerjayaOne PLT and themed "Learning from the Past, Looking to the Future," the gathering convened senior leaders from corporate and institutional sectors to address a significant leadership transition challenge. Datuk Seri Abdul Mutalib Razak, chairman of KerjayaOne, characterised wisdom as something that emerges through lived experience—through making difficult choices under uncertainty, managing organisational crises, learning from both triumphs and failures. This definition stands in sharp contrast to AI's capacity to process information. Wisdom involves integrating knowledge with judgment, experience with ethics, and technical expertise with institutional values in ways that remain distinctly human enterprises.
The conference brought together prominent figures including Tan Sri Che Khalib Mohamad Noh, former president of Tenaga Nasional Berhad, alongside Tan Sri Abdul Wahid Omar, who chairs both Cypark Resources Berhad and IOI Corporation Berhad, and Datuk Dr Mohd Khalis Abdul Rahim, former chief human capital officer at Malaysia Aviation Group. Their participation underscored recognition across Malaysia's corporate and institutional leadership that institutional knowledge and mentorship cannot be digitised or algorithmically replicated. The succession challenges facing Malaysian organisations require deliberate mechanisms for transferring experiential wisdom from retiring leaders to their successors, processes that inherently depend on personal engagement and human relationship.
For Malaysian policymakers and institutional leaders, Tun Zaki's warnings carry particular salience as the nation navigates technological transformation. The temptation to defer complex institutional decisions to artificial intelligence systems—in pursuit of efficiency, consistency, or perceived objectivity—poses genuine risks. While Malaysia should certainly embrace AI's legitimate applications in data analysis, case management, and preliminary research, the fundamental governance and judicial functions must remain under human stewardship. This requires establishing clear regulatory frameworks, as the DIFC did, that govern AI use in sensitive contexts while preserving transparency and accountability.
The path forward involves neither wholesale rejection of artificial intelligence nor uncritical embrace of technological solutionism. Instead, Malaysian institutions must develop thoughtful integration strategies that treat AI as a powerful subordinate tool serving human decision-makers, not as a replacement for them. This approach respects both the genuine capabilities that AI brings to modern governance and the irreducible importance of human wisdom, ethical reasoning, and institutional judgment. As Malaysia continues developing its digital economy and governance infrastructure, maintaining this principled stance—one that Tun Zaki has articulated with clarity—will prove essential to ensuring that technological progress enhances rather than undermines the quality of institutional leadership and judicial integrity.
