Artificial intelligence will continue to transform how institutions operate, but it can never fully supplant the irreplaceable element of human judgment that underpins effective governance and judicial decision-making. This was the central message delivered by former Chief Justice Tun Zaki Azmi at the Wisdom Malaysia 2026 Conference in Kuala Lumpur, where he emphasised that while AI offers valuable analytical capabilities, the ultimate responsibility for critical decisions must remain squarely with humans who possess accumulated wisdom and ethical grounding.
Tun Zaki, currently chairman of Pertama Digital Berhad, articulated a nuanced position on technology's role in the justice system. He acknowledged that AI can enhance preliminary research, offer fresh analytical perspectives, and streamline routine aspects of legal work. However, he stressed that judges cannot delegate their core function of interpreting legislation and applying it to specific cases. The risk of AI systems producing plausible-sounding but factually incorrect outputs—a phenomenon known as "hallucinations"—means that blind reliance on automated analysis would fundamentally compromise judicial integrity.
Drawing upon his extensive experience as former Chief Justice of the Dubai International Financial Centre (DIFC) Courts, Tun Zaki highlighted a practical safeguard his institution implemented to manage AI integration responsibly. The DIFC established a mandatory disclosure requirement obligating lawyers and legal researchers to notify the court whenever artificial intelligence has been used in preparing legal arguments or court documents. This transparency mechanism serves a dual purpose: it ensures judicial officers remain alert to potential vulnerabilities in AI-generated materials, and it creates an auditable record of how technology influences case preparation.
The former Chief Justice's arguments gain particular resonance when examined through the structure of appellate proceedings. He explained that multi-judge panels with odd numbers—typically three, five, seven, or nine justices—exist precisely because human reasoning is inherently variable. Different judges, despite possessing similar qualifications and access to identical evidence, may interpret statutory language differently or assess witness credibility through divergent lenses. This diversity of judicial perspective, rather than being a weakness, represents a strength that protects the system from individual error or bias. An algorithm, by contrast, would produce uniform outputs regardless of whether those outputs reflected sound legal reasoning or embedded biases in its training data.
For Malaysia's legal and institutional framework, this discussion carries immediate practical implications. As Malaysian courts and government agencies increasingly explore AI applications in document review, case prediction, and administrative efficiency, Tun Zaki's warnings suggest the need for careful policy development. The DIFC model of mandatory AI disclosure could serve as a template for Malaysian courts, ensuring that judicial officers retain meaningful oversight of technology-assisted legal research. Without such safeguards, the judiciary risks ceding analytical control to systems that lack accountability and interpretative authority.
Tun Zaki further articulated that wisdom accumulates through navigating genuine complexity, making consequential decisions under uncertainty, and reflecting upon both successes and failures. These experiential dimensions cannot be replicated by algorithms processing historical data. A human judge who has seen how legal principles apply across decades of cases develops intuitive understanding that transcends what explicit rules or statistical patterns can capture. Similarly, leaders who have stewarded organisations through crises develop judgment that no computational model—however sophisticated—can fully replicate.
This perspective aligns with the conference's broader theme of intergenerational knowledge transfer. Datuk Seri Abdul Mutalib Razak, chairman of KerjayaOne PLT, which organised the two-day gathering, framed the event as an intentional effort to preserve institutional wisdom before it departs with retiring leaders. By assembling experienced figures—including former Tenaga Nasional Berhad president Tan Sri Che Khalib Mohamad Noh, Cypark Resources Berhad and IOI Corporation Berhad chairman Tan Sri Abdul Wahid Omar, and former Malaysia Aviation Group chief human capital officer Datuk Dr Mohd Khalis Abdul Rahim—the conference created space for leaders of different generations to engage in knowledge exchange that cannot be codified into algorithms or databases.
The challenge Malaysia and other Southeast Asian nations face involves calibrating AI adoption appropriately. In certain domains, such as initial document categorisation, regulatory compliance screening, or resource allocation analysis, AI can genuinely enhance efficiency without compromising core functions. However, deploying AI in domains requiring normative judgment, credibility assessment, or interpretation of ambiguous rules risks degrading institutional quality. The judiciary, in particular, must resist pressure to adopt technologies that promise speed at the cost of careful deliberation.
Tun Zaki's warnings also resonate within the broader context of digital governance rollouts across the region. As governments implement e-services, automated decision-making systems, and AI-powered administrative processes, the principle he articulated—that technology must augment rather than replace human judgment—deserves central consideration. Public trust in institutions depends partly on the perception that consequential decisions reflect human reasoning, ethical consideration, and accountability mechanisms that citizens can understand and challenge.
The requirement that lawyers disclose AI usage in the DIFC Courts creates an elegant solution that manages innovation while preserving oversight. Malaysian courts might similarly establish disclosure protocols, perhaps with accompanying guidelines about when AI assistance is permissible in litigation and when judges should exercise heightened scrutiny over AI-generated materials. Such frameworks need not stifle technological adoption but can channel it toward legitimate efficiency gains while protecting the human elements essential to justice.
Ultimately, Tun Zaki's message reflects accumulated institutional wisdom about technology's proper place. AI excels at tasks involving pattern recognition, speed, and consistency. It falters in domains requiring contextual sensitivity, ethical reasoning, and accountability to affected parties. The judicial system, which exists to serve justice for individuals facing life-altering consequences, cannot afford to subordinate these human-centric requirements to technological efficiency. The conference's intergenerational dialogue, paired with measured warnings about AI's limitations, offers Malaysian institutions a valuable opportunity to develop governance frameworks that harness technology's benefits while protecting the irreducible role of human judgment.
