Artificial intelligence, despite its rapid advancement and growing capabilities, cannot and should not displace human judgment in critical decision-making processes, according to Tun Zaki Azmi, former Chief Justice of Malaysia. Speaking at the Wisdom Malaysia 2026 Conference in Kuala Lumpur on August 5, Zaki, who currently chairs Pertama Digital Berhad, argued that AI functions best when deployed as an analytical instrument that enhances human deliberation rather than substitutes for it. His cautionary stance reflects a broader institutional concern about the unchecked adoption of algorithmic systems in sectors where accountability, ethical reasoning, and nuanced interpretation remain fundamentally human responsibilities.
The distinction Zaki draws between AI as an assistive tool and AI as a replacement is critical for understanding how technology should integrate into Malaysia's legal and institutional frameworks. While artificial intelligence excels at processing vast datasets, identifying patterns, and generating preliminary analyses with speed and consistency, it lacks the experiential foundation and moral grounding necessary for consequential decisions. In judicial contexts particularly, the application of law requires judges to wrestle with ambiguity, precedent, societal norms, and the unique circumstances of individual cases—functions that demand the interpretive capacity that only trained human minds can reliably execute. Zaki's emphasis on this limitation reflects decades of judicial practice and exposure to technological implementation across different legal jurisdictions.
Drawing on his extensive experience as Chief Justice of the Dubai International Financial Centre (DIFC) Courts, Zaki highlighted a concrete institutional response to AI integration. The DIFC established a directive requiring lawyers and legal researchers to explicitly disclose whether artificial intelligence tools have been used in preparing legal documents, research materials, or court submissions. This transparency mechanism serves as a critical safeguard, enabling judges to exercise heightened scrutiny over AI-generated content and maintain oversight of the reasoning processes underlying legal arguments. For Malaysian courts and institutional bodies considering similar adoption of AI tools, the DIFC model offers a pragmatic framework that acknowledges technology's utility while institutionalizing accountability and human review at pivotal decision points.
A particular vulnerability Zaki identified is AI's propensity for what technologists term "hallucinations"—instances where machine learning systems generate plausible but entirely fabricated information, citations, or reasoning chains that lack any factual basis. In judicial proceedings, such errors could catastrophically undermine legal arguments or case analysis if deployed without rigorous human verification. Judges who rely on preliminary AI analysis must therefore treat such outputs as preliminary scaffolding rather than substantive findings, subjecting them to the same critical scrutiny and independent verification they would apply to any contested factual or legal assertion. This requirement for supplementary human validation fundamentally preserves the judge's role as the final arbiter of fact and law.
Zaki's broader argument rests on observing how appellate courts institutionalize dissent and multiple perspectives. The Federal Court, Court of Appeal, and other higher courts sit in panels of odd-numbered judges—three, five, seven, or nine—specifically because human beings legitimately disagree about how to interpret law, assess evidence credibility, and apply precedent to novel circumstances. Different judges, drawing on varied experiences and judicial philosophies, may reach divergent conclusions from the same factual record. This institutional pluralism is not a bug but a feature; it ensures that judicial decision-making benefits from multiple viewpoints and that majority opinions emerge only after rigorous deliberation. No algorithm, regardless of sophistication, can replicate this deliberative process or the legitimacy that flows from reasoned disagreement among qualified jurists.
The conference itself, organized by KerjayaOne PLT and themed "Learning from the Past, Looking to the Future," underscores the original impetus driving Zaki's intervention. KerjayaOne chairman Datuk Seri Abdul Mutalib Razak framed the gathering as an intergenerational knowledge-transfer initiative, designed to preserve institutional wisdom accumulated through decades of crisis management, ethical decision-making under pressure, and learning from both triumph and failure. In this context, Zaki's remarks function not merely as a technical caution about AI limitations but as a philosophical statement about the irreplaceable value of human experience and judgment in organizational and institutional leadership. Wisdom, by definition, cannot be algorithmically generated; it emerges through lived experience, mentorship, reflection on failure, and the integration of diverse perspectives over time.
For Malaysia's judicial system and broader institutional ecosystem, Zaki's message carries immediate practical implications. As courts and government agencies explore AI applications to improve efficiency—from case management systems to legal research databases—the imperative is to establish clear governance frameworks that subordinate technological deployment to human oversight. This might entail mandatory disclosure requirements similar to those at the DIFC, regular audits of AI-generated preliminary findings, and periodic retraining of judicial and institutional staff to ensure they maintain critical distance from algorithmic outputs. Such measures are not obstacles to technological progress; rather, they represent the institutional scaffolding necessary to harness AI's genuine benefits while preserving the human judgment upon which rule of law and institutional legitimacy ultimately depend.
The timing of Zaki's intervention is significant given Malaysia's ongoing digital transformation agenda and the enthusiastic adoption of AI across multiple sectors. While business and administrative contexts may tolerate higher levels of algorithmic autonomy, the judiciary and other truth-determining institutions require different standards. Courts do not merely process information; they authoritatively resolve disputes, determine culpability, and allocate rights and penalties. These functions carry consequences that ripple through individuals' lives and the broader social fabric. Allowing algorithmic systems to operate without meaningful human review in such domains would represent a fundamental abdication of institutional responsibility and would erode public confidence in the impartiality and reliability of judicial outcomes.
The broader ecosystem of leaders participating in the Wisdom Malaysia 2026 Conference—including figures such as 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—collectively represents decades of accumulated institutional knowledge across corporate, utility, and organizational leadership domains. These leaders have navigated technological disruption, competitive pressures, and organizational crises through judgment calls that no algorithmic system could have executed with equivalent nuance and contextual sophistication. By convening such figures for intergenerational dialogue, the conference acknowledges that the primary challenge facing Malaysian institutions is not the lack of data or analytical horsepower but the preservation and transfer of judgment capability across generational transitions.
Zaki's final assertion—that AI will make human work easier but will not replace human minds—encapsulates a measured optimism about technological integration grounded in institutional realism. The judicial and institutional communities need not reject artificial intelligence outright; indeed, properly deployed tools can reduce administrative burden, accelerate preliminary analysis, and free experienced professionals to focus on aspects of their work requiring genuine judgment. However, this instrumentalist approach demands institutional discipline and transparency. Courts and institutions must resist the temptation to outsource accountability to algorithmic systems or to treat AI outputs as authoritative findings merely because they are machine-generated. The final responsibility for interpretation, reasoning, and decision-making must remain with accountable human agents—judges, executives, leaders—whose authority derives from expertise, ethical commitment, and democratic legitimacy rather than from technological capacity.
