How HR Can Ensure Fairness and Accountability in AI-Powered Dismissals Under Singapore’s New Laws

AI-Driven Dismissals: What HR Must Get Right Under Singapore’s New Fairness Laws

As AI systems become integral to performance evaluations and workforce planning, employers in Singapore face heightened responsibility. When AI influences employee terminations, the risk of errors rises sharply. Although legislation is still evolving, HR’s duty to ensure fairness, transparency, and sound judgment is clear.

This article outlines key points HR leaders must address when using AI tools in employee decision-making, especially with the upcoming Workplace Fairness Act 2025 (WFA) on the horizon.

Applying Tripartite Standards to AI-Linked Dismissals

AI can improve decision consistency but does not replace the legal need for fairness. Terminations must be based on clear, merit-driven grounds—not unverified misconduct or discriminatory reasons. This aligns with Singapore’s Tripartite Guidelines on Fair Employment Practices and the WFA, which prohibits discrimination based on 11 protected characteristics.

Transparency is critical. Employers should clearly explain how AI models work, the data used for training, and the checks in place against bias. Most importantly, human judgment must remain central. At least one person should validate the dismissal decision, and the employee must have the opportunity to respond.

If AI Flags ‘Poor Performance,’ Is That Enough?

While Singapore law does not require employers to provide reasons for termination, relying solely on AI-generated flags is risky—especially if the employer cannot explain the AI’s output. Without a clear, job-relevant basis, wrongful dismissal claims may arise.

Employers need to establish clear performance benchmarks and ensure managers can articulate how AI influenced decisions. Treat the AI system as an explainable tool, not a black box. The logic behind AI outputs must be documented and understandable, particularly in termination records.

How Transparency Influences Legal Risk

AI promises objectivity, but a lack of transparency can expose employers to legal challenges. When decisions are attributed to AI, employers must explain why and how the tool affected outcomes.

Inconsistencies in rationale or poor documentation weaken legal defenses. Even if the law sets minimal disclosure standards, transparency backed by records is a crucial safeguard.

Accountability for AI Vendor Tools

Many companies outsource AI-based appraisal and selection tools, but employers remain fully responsible for HR decisions supported by these systems. They must understand how AI works and ensure outputs are free from indirect discrimination.

Algorithms can inadvertently factor in protected traits like accents or educational background. Employers should demand full transparency from vendors about what the AI measures, how it processes data, and what bias mitigation mechanisms exist. Without this, defending decisions is difficult.

AI Monitoring: Obligation to Inform Staff

Singapore’s Personal Data Protection Act (PDPA) allows employers to use employee data for performance evaluation without consent under an evaluative purpose exception. However, there remains an obligation to inform employees when AI is used to monitor performance.

Employers should disclose the purpose and scope of AI monitoring, especially if it impacts employment decisions. Updates to employee handbooks should clearly state how data is collected and used in performance appraisals.

Human-In-The-Loop (HITL) Processes

Superficial AI review won’t hold up in sensitive cases like dismissals. Meaningful human oversight is essential. The reviewer must have authority to override AI and ensure fairness.

Define who reviews AI outputs, at what stage, and with what powers. Internal policies should specify which HR decisions need human review, how AI results are validated, and how employees can report concerns. In termination or redundancy cases, human validation is a key defense.

Documenting AI-Influenced Decisions

Good documentation supports internal governance and legal readiness. Keep clear, dated records of:

• The AI tools used
• How employee data was assessed
• Who reviewed and approved the decision
• Whether the employee was given a chance to respond

Short internal memos or emails can prove the process was consistent and thoughtful. Vague or shifting explanations leave companies vulnerable to claims.

Appeal Mechanisms in AI-Backed HR Systems

With the WFA approaching, establishing internal appeal channels is vital, especially for AI-influenced decisions. Employees need a way to raise concerns about fairness or data accuracy.

Appeals should be handled by a separate HR contact or panel. Listening to employees not only addresses grievances but also helps identify system weaknesses early.

What Happens if AI Tools Are Misused?

Reported cases of AI misuse in termination are still few in Asia, but risks are rising. For example, an AI algorithm selecting redundancies based on attendance, tenure, and salary could unintentionally discriminate against older workers or caregivers.

The key issue is whether employers can explain and justify AI decisions. Lack of transparency or flawed training data opens the door to discrimination claims.

What Should HR Implement Now?

The WFA will be enforced by 2026 or 2027, but HR teams should act immediately. Implement internal AI policies, solid HITL review processes, and clear appeal mechanisms covering the entire employment lifecycle.

Understanding AI tools deeply is critical. Without insight into how vendors’ systems operate, employers cannot defend decisions effectively.

AI can speed up assessments and organize data neatly, but without accountability, it introduces risk. When performance scores affect careers, HR must ensure processes are fair, transparent, and defensible.

With proper safeguards—clear documentation, real-time oversight, and structured appeals—AI can support better decision-making. Expectations are shifting even before the law catches up.