COVID-19 Update

COVID-19: Impact on SSHRC programs, experts database and perspectives from our community.

Algorithmic HR management: Identifying best practices and worker impacts

Printable version

About the project

Any technological advance implemented in the workplace will cause changes to workers’ duties and how they are done. Current technologies, especially ones powered by artificial intelligence (AI), are no exception. Thanks to their considerable advantages, AI applications are rapidly deploying in a wide variety of organizational activities and have already brought about significant changes to the nature of many jobs. Furthermore, advances in AI-based algorithms allow for the transformation or automation of managers’ duties as well as those of workers under them. The increased use of algorithms in management tasks, including employee management, is referred to as “algorithmic management.”

Since the first studies on the subject were published in 2015, algorithmic management has been the focus of increasing attention from researchers. This is primarily due to the swift spread of algorithmic management, but also because of the significant issues it raises regarding many aspects of quality of life at work. In this fast‑developing research, it is important to take stock of current knowledge in order to identify the most promising approaches for future work and to provide early science‑based observations to guide organizational practices. This report presents the conclusions drawn from a review of the empirical studies on algorithmic management published to date.

Key findings

  • The HR management activities most commonly performed wholly or partly by algorithmic systems are: monitoring of a multitude of elements associated with work activities, work planning (assigning tasks, setting goals and schedules), performance evaluation and feedback, compensation calculations, and employment contract dismissal or termination.
  • Generally speaking, studies show more negative or adverse effects on workers’ quality of life connected to the use of algorithmic management. Fortunately, however, our findings also enabled us to identify certain factors (system characteristics and management practices) that help to mitigate or cancel out those impacts.
  • The presence or perceived degree of presence of algorithmic management seems to be associated with a number of emotions, attitudes and behaviours from workers, such as anxiety, resistance, trust and mistrust, a perception that a psychological contract between the organization and its employees has been broken, low work engagement, frustration, worry, decreased motivation, perceived injustice and/or feelings of dehumanization at work.
  • The presence or perceived degree of presence of algorithmic management seems to correlate with certain working conditions and ways to organize work. The studies we reviewed found that algorithmic management can simultaneously encourage collective action for resistance or mutual assistance, increase competition among workers and/or decrease support between colleagues. It is also associated with an increased workload, power asymmetries between the organization and its workers, increases in requirements for requalification or professional development, and reduced independence for workers. It also poses occupational health and safety risks as well as risks for work‑life balance, and may potentially be associated with an increase in systematic discrimination and greater job and financial precarity.
  • The effects of algorithmic management on organizations at large have received less attention from researchers. However, the findings of some studies have implied that it could result in services that are faster, more efficient and more standardized, and could help optimize production costs. Further research is needed to clarify the longer‑term effects of algorithmic management on organizational performance.

Policy implications

  • These conclusions support all use of initiatives intended to improve the responsible use of algorithmic management in order to support workers’ dignity and well‑being.
  • The studies we reviewed showed that a number of parameters and characteristics of algorithmic systems contribute to mitigating the risks mentioned above. These include transparency about the existence of such a system and how it works, about the reliability and accuracy of its decisions, about how fair the system and its decisions are, and about how much power is held by employees and algorithms respectively.
  • Our findings also show that it is important to safeguard workers’ independence in carrying out their duties, and to put more weight on trust than on control.
  • These conclusions indirectly suggest that organizational self‑regulation of algorithmic management systems has proven insufficient, and that this phenomenon would warrant a legal framework that lays out employer obligations in terms of governance and workers’ rights.

Further information

Read the full report

Contact the researchers

Xavier Parent-Rocheleau, Assistant Professor, Department of Human Resources Management, HEC Montréal:

Marie-Claude Gaudet, Assistant Professor, Department of Human Resources Management, HEC Montréal:

Marylène Gagné, Professor, Future of Work Institute, Curtin University:

Pamela Lirio, Associate Professor, School of Industrial Relations, Université de Montréal:

Antoine Bujold, PhD student, HEC Montréal:

The views expressed in this evidence brief are those of the authors and not those of SSHRC, the Future Skills Centre or the Government of Canada.

Date modified: