As artificial intelligence becomes a near-default tool in hiring, new research warns that companies may be oversimplifying what “fairness” really means, the Harvard Business Review writes.
A three-year field study of a global consumer goods company found that AI systems designed to reduce bias often lock in a single, rigid definition of fairness—typically consistency—while sidelining other valid perspectives, such as local context and managerial judgment. By replacing résumé reviews with AI-driven assessments, the company improved standardization but gradually narrowed its candidate pool and reduced flexibility in hiring decisions.
The research argues that fairness is not something AI delivers automatically at launch. Instead, it is shaped by the people who design, deploy and defend these systems.
Leaders are urged to ask tougher questions about who defines fairness, which values get encoded into algorithms and what perspectives are quietly excluded over time. Without ongoing oversight and debate, well-intentioned AI tools can unintentionally entrench bias rather than eliminate it.
Read the full story. A subscription may be required.
