AI sees patterns, humans see people: Restoring judgment to HR decision-making
- Josephine Tan
In the modern corporate landscape, AI is no longer just a productivity tool; it is a decision-maker. From CV screening to performance rewards and promotion modelling, AI is rewriting the rules of organisational logic. However, as Rita Tsui, Founder of AsiaHRM, recently highlighted in a LinkedIn Live session for the Sustainability in Business Series, this shift brings a critical risk known as the “black box” effect. Tsui emphasised that governance is everyone’s responsibility, drawing parallels to historic corporate collapses like Enron and Lehman Brothers to illustrate that poor oversight directly threatens business sustainability.
Tsui illustrated the stakes through a tale of two leadership groups evaluating identical candidates. In the first group, leaders openly debated assumptions and provided the unsuccessful candidate with clear, aligned feedback, leaving the individual with their respect for the brand intact despite not getting the job. In contrast, the second group relied on a system-generated recommendation with minimal human discussion. This resulted in a frustrated candidate who felt processed by an opaque system and vowed never to apply again. The core difference was not the ultimate decision, but the governance behind it; one group treated decision-making as a human narrative that required explanation, while the other treated it as a mere computation.
For HR professionals, the risks of ungoverned AI are manifold and potentially damaging. One primary concern is scaled bias, where AI learns from historical data and repeats past inequities—such as favouring specific genders or backgrounds—faster and more consistently than a human ever could. Additionally, an “accountability gap” often arises when organisations mistakenly assume that a vendor or algorithm is responsible for errors. Tsui was firm in her stance that AI cannot be held accountable, and that only humans can be responsible for checking data and validating models. Without clear documentation, an organisation cannot justify, defend, or improve its decision-making processes.
To help organisations navigate this transition, Tsui outlined a five-level maturity scale to measure how future-ready an HR department truly is. At level one, decisions are ad hoc and depend entirely on individual managers. Level two introduces defined rules that are often ignored, while level three achieves consistency through standardised processes. Level four involves regular audits and fairness checks to detect bias early. Finally, level five represents the gold standard of ethical AI, where governance is proactive and includes “human-in-the-loop” oversight to ensure transparency by design.
READ MORE: The future of digital work takes shape at HR Tech Asia 2026
The goal for HR leaders is to move beyond seeing governance as a mere compliance exercise and instead view it as a core business strategy. This requires a robust structure to provide clarity, a transparent process to ensure consistency, and a culture that encourages leaders to model accountability. Tsui concluded that while structure gives an organisation its rules, culture is what gives it trust. By making governance a strategic priority, HR can become the backbone of an organisation that is fair, transparent, and trustworthy in an increasingly automated world.
To dive deeper into these strategies, HR professionals can attend Tsui’s upcoming presentation at HR Tech Asia 2026 in Singapore. Her session, titled Association Spotlight: HR Governance in the Age of AI: Building Fair, Transparent, and Trustworthy People Systems, will take place on May 5 at the Power Talk Stage. The presentation will offer insights into ensuring ethical use of employee data, defining HR’s role in AI oversight, and using governance to strengthen organisational culture. Registration is currently open for those looking to lead their workforce into the AI era with confidence.
