Artificial intelligence (AI) has the potential to fundamentally transform the world of work. Particularly in the context of human resources management, the use of AI-based systems is increasingly becoming a focal point for public debate. Because in addition to the oft-cited opportunities for optimising efficiency, these systems come with a number of significant risks, especially due to their potential for employee surveillance.
Against the backdrop of the COVID pandemic, AI in human resource management gains momentum
Current debates on AI in human resources management are driven by three key factors.
- Technological progress and the increasing availability of data. Digital collaboration during the pandemic fueled the availability of large amounts of new behavioural data of employees.
- New challenges in human resource management, with AI applications promising to compensate for the loss of control that managers experience as a result of the increase in remote working. The applications also offer employees personalised feedback, helping them manage themselves when working from home.
- Recent attempts to regulate, such as the EU’s proposed AI regulation that classifies AI in employment law, and therefore also in human resource management, as high-risk applications.
What lies behind the trend?
Even though academic studies on how widespread such systems are currently being used remain rare, particularly in Germany, the large number of potential use cases shows that they can be used at every stage of the employee life cycle. They range range from recruitment algorithms to productivity analyses to predicting how likely an employee is to resign. Applications for automated human resource management are frequently summarized under the umbrella term “People Analytics” (PA). They increasingly include functions based on complex statistical models, often referred to as artificial intelligence. These applications collect information on the behaviour and characteristics of employees, often combine this with external data, and use descriptive and predictive analyses to evaluate them – thereby providing actionable insights on employee performance and how to optimise it.
The way human users interact with and use such applications is often characterised by (overly) high expectations
At the same time, these systems are clearly ambivalent: on the one hand offering the immense potential for surveillance, and on the other the potential for supporting employees. They also call into question the role of managers: no human decision-maker will ever be able to match these systems in terms of the speed and quantity of data processing. Nevertheless, the potential for optimising and objectifying decision-making is often exaggerated. This is based on three main assumptions: that complex human activities and qualities can be quantified and measured in the workplace; that human behaviour can be predicted using historical data; and that algorithms are superior to human decision-makers.
There is currently a lack of empirical evidence on the risks and opportunities this creates for the digital world of work
Although there is a long tradition in management research of questioning the extent to which complex workplaces can be controlled, such questions are being asked anew in light of the developments outlined above. For practitioners too, data-driven management has dominated the lists of human resources trends for years. However, much of the existing research is conceptual and reveals an urgent need for empirical research – especially on the particularities of the German market. The project by Research Fellow Sonja Köhne aims to help fill this gap, because many experts and researchers agree that the importance of data in human resources management will increase further. Currently, however, we still know too little about what is already being used in practice in Germany and how this will affect relevant actors in organisations and their work practices.
What is the objective of the project?
The project aims to compare and contrast the offers currently available and their meaning for relevant corporate actors in Germany. A qualitative research approach will be used to examine available applications, the extent to which relevant corporate actors are willing to use them and how they affect work practices. The project therefore consists of two phases:
- First, the available applications and their value propositions will be examined using a market analysis and expert interviews.
- Second, the project will analyse the meaning of these applications for relevant corporate actors through qualitative interviews with human resources managers, people managers, employees and worker representatives.
The project therefore ties in closely with previous work of the Policy Lab Digital, Work & Society, and generates insights on the risks and opportunities of AI-based systems in people management. It will also outline potential human-centric ways in which such systems could be developed for the Germany context, with particular regard to the prevailing legal framework in areas such as data protection.
- The Fellowship Programme
- Employee Data Protection Advisory Committee
- Sonja Köhne’s profile at the Alexander von Humboldt Institute for Internet and Society (HIIG)
- re:publica video: People Analytics In The Workplace: Every Break You Take, Every Click You Make
- Blog Post: Working from Home but Never Alone: Why People Analytics Have to Be Designed with the Employee in Mind