Around the world, the conversation about Artificial Intelligence in public procurement is dominated by the promise of efficiency. The focus is on streamlining processes, automating tasks, and achieving significant cost savings. Studies, such as a recent one by Boston Consulting Group, project remarkable outcomes like up to 15% in savings and a significant reduction in human workload. Yet, in our Malaysian context, to focus solely on these benefits would be to miss a far more critical opportunity: leveraging AI as a frontline tool in the battle against corruption.

The timing could not be more urgent. The recent MACC revelation that Malaysia lost RM277 billion over six years, much of it through collusion in public tenders, is a stark reminder of the deep-seated challenge we face. As we grapple with this reality, the small nation of Albania has embarked on a controversial experiment. Faced with its own entrenched corruption, its government has appointed an AI digital assistant to oversee its entire public procurement process, hoping to create a system free of human bias and graft—a move now facing intense scrutiny from technical and legal experts.

The potential benefits of deploying such technology in Malaysia are immense. Imagine an AI system as an incorruptible digital auditor, capable of analyzing thousands of bids simultaneously. It could flag suspicious patterns invisible to the human eye—interconnected companies winning contracts repeatedly or bids that are consistently just below the threshold for extra scrutiny. By ensuring every decision is data-driven and transparent, we could theoretically restore fairness, save billions in public funds, and begin to rebuild the deep deficit of public trust.

However, recent developments show we must proceed with extreme caution. Experts are now questioning the entire premise of an “incorruptible” AI, pointing out that any system is only as good as the data it is fed. As one political scientist warned, if a corrupt system provides manipulated data, the AI will merely “legitimise old corruption with new software.” This also raises a critical question of accountability—an issue so serious it is being challenged in Albania’s Constitutional Court. If a machine makes a flawed decision, who is responsible?

The most prudent path for Malaysia, therefore, is likely not the appointment of a full “AI minister.” Instead, we should explore a more pragmatic, hybrid model. Let us envision AI not as a replacement for human decision-makers, but as a powerful, mandatory tool to support them. Our MACC, government auditors, and procurement boards could be equipped with AI systems designed to act as a first line of defense. This “digital watchdog” could flag high-risk tenders for stringent human review, catching cases that might otherwise be missed due to simple human oversight or inherent bias. Furthermore, its data-driven recommendations would serve as objective evidence of impartiality, making it much harder for legitimate cases to be dismissed due to personal or political agendas.

The unfolding experiment in Albania, with all its emerging challenges, has opened a vital, global conversation. For a nation like ours, which has lost so much to this long-standing problem, ignoring the potential of technology to enforce integrity is no longer an option. It is time to seriously innovate our way towards better governance.