BY PRISCILLA JEBARATNAM

Artificial intelligence (AI) is rapidly transforming workplace practices, including how employee performance is assessed. Organizations are increasingly adopting AI-enabled evaluation systems because they promise greater consistency, efficiency, and objectivity. But an important question remains: Do more accurate evaluations automatically lead employees to perceive them as fair?

A recent rapid evidence assessment examining nine empirical studies published between 2019 and 2025 suggests the answer is more complicated than many organizations assume.

Accuracy and Fairness Are Not the Same

Across the studies reviewed, AI-driven performance evaluation systems consistently improved technical aspects of assessment. These systems demonstrated stronger consistency, reliability, and predictive accuracy while reducing some forms of human subjectivity.

However, better technical performance did not automatically translate into greater employee trust or perceptions of fairness.

Employees appeared to evaluate AI systems not only by the outcomes they produced but also by how those outcomes were generated. In other words, fairness depends as much on the process as on the results.

Governance Matters More Than Technology

One of the strongest findings from the review was that employees’ perceptions of fairness were shaped primarily by governance practices rather than by algorithmic performance itself.

Four factors consistently emerged as critical:

• Transparency – Employees want to understand how AI-generated decisions are made.
• Explainability – Workers are more likely to trust systems when they can see and question the reasoning behind evaluations.
• Ethical data governance – Confidence increases when organizations use job-relevant and representative data while minimizing bias.
• Employee involvement – Participation in system design and implementation strengthens trust and acceptance.

When these elements were present, employees were more likely to view AI evaluations as legitimate and procedurally fair.

The Importance of Explainability

Many AI systems operate as “black boxes,” producing recommendations without clearly showing how conclusions are reached. The review found that such opacity often increases concerns about bias and reduces trust.

Conversely, explainable AI serves as a bridge between technical performance and employee acceptance. Employees do not necessarily need to understand every algorithmic detail, but they want assurance that evaluation processes are understandable, accountable, and open to scrutiny.

This finding may be particularly important in higher education and other knowledge-based environments where autonomy, professional judgment, and shared governance are highly valued.

Data Quality Shapes Legitimacy

Employees also pay attention to the information used to evaluate them.

Systems built on job-related variables and representative datasets were viewed more favorably. In contrast, the inclusion of demographic characteristics or other non-performance-related factors often raised concerns about discrimination, privacy, and fairness.

These findings highlight that trust in AI is influenced not only by outcomes but also by perceptions regarding the appropriateness of the underlying data.

Human Involvement Still Matters

The review further suggests that fully automated performance evaluations may not represent the optimal approach.

Organizations that involved employees in implementation, encouraged feedback, and maintained human oversight reported stronger levels of trust and acceptance. By contrast, highly surveillance-oriented systems often reduced perceptions of autonomy and fairness, even when employees acknowledged the technical capabilities of the technology.

These findings support the value of hybrid human–AI models that combine algorithmic efficiency with managerial judgment and interpersonal communication.

Looking Ahead

As AI continues to reshape performance management, organizations should recognize that technical accuracy alone is insufficient. Effective AI systems must be embedded within transparent, participatory, and ethically governed environments.

Ultimately, successful AI adoption is not simply a technological challenge—it is an organizational and human one. Building trust requires more than sophisticated algorithms. It requires designing systems that employees perceive as fair, understandable, and aligned with the values of the workplace.

When organizations prioritize transparency, explainability, ethical governance, and employee voice, AI has the potential to improve not only evaluation accuracy but also organizational trust and legitimacy.

Priscilla Jebaratnam is a doctor of business administration candidate at the University of Maryland Global Campus with more than twenty years of experience in financial services, higher education, and organizational operations. Her research focuses on the ethical implementation of artificial intelligence in employee performance evaluation, with an emphasis on fairness, transparency, and organizational justice in online higher education. Through her work, she aims to bridge academic research and practical application by providing evidence-based insights that help organizations implement AI responsibly and effectively.