The International Labour Organization (ILO) has warned that companies worldwide are adopting Artificial Intelligence (AI) tools in Human Resource Management without fully understanding their limitations, leading to serious risks for workers and employers.
The warning comes from a new working paper published in November 2025 in Geneva. The report, authored by senior economist Dr. Janine Berg and HRM expert Ms. Hannah Johnston, highlights that AI is rapidly permeating recruitment, compensation, scheduling and performance management, yet often produces poor results due to weak objectives, low-quality data and opaque programming.
Releasing the paper, the authors said the study was designed to help HR managers understand why AI systems fail, how failures occur and what HR professionals must do to prevent legal, ethical and operational challenges. They stressed that HR departments, eager to appear more “data-driven,” often deploy AI tools that expose workers to discrimination, flawed decision-making and inaccurate assessments.
The report, titled AI in Human Resource Management: The Limits of Empiricism, argues that HR leaders increasingly embrace AI because they want their decisions to appear more scientific. Dr. Berg warned that this belief is misguided.
“AI tools often reduce complex human behaviour into simple numbers that do not represent reality,” she said.
Ms. Johnston noted that HR managers frequently adopt AI systems built by third-party vendors despite not understanding how the tools reach their conclusions. Many rely on biased or poor-quality data, she said, resulting in discrimination in job advertisements, CV screening and salary recommendations.
Recruitment, according to the report, is the HR function most affected by AI misuse. The authors cited cases in which algorithms displayed job ads primarily to men because they were programmed to minimise advertising costs rather than ensure fairness. Dr. Berg said this demonstrated how AI “learns the wrong lessons” when objectives are poorly defined.
The paper also highlighted emerging risks in pay-setting, where algorithms personalise wages based on perceived worker behaviour or acceptance trends. Ms. Johnston warned that this could lead to “algorithmic wage discrimination,” particularly affecting women and historically disadvantaged groups.
AI-driven scheduling systems were found to improve efficiency but often ignore human constraints, contributing to stressful working conditions. Similarly, AI-powered performance management tools—particularly those analysing facial expressions, tone or computer activity—were flagged for generating misleading or intrusive feedback.
The authors urged businesses to slow down and scrutinise AI systems thoroughly, including how objectives are defined, how data is sourced and how algorithms are programmed. They advised HR departments to participate in system design rather than relying on ready-made products.
Dr. Berg added: “HR leaders must understand AI before trusting it. Failure to do so could harm workers and damage organisational reputation.”
The ILO said the paper aims to guide HR executives, policymakers and organisations in adopting AI responsibly while safeguarding fairness, job quality and compliance with labour standards.
