Please note: Most of the papers and findings of our research project “Automated Human Resources Management and Labor Rights” are currently available only in German (German web dossier). The PDF file of the ethical anaylsis People Analytics must benefit the people. An ethical analysis of data-driven algorithmic systems in human resources management can be found here.
Companies use such systems to identify employees to retain, to support, or to promote. Which key member of staff is likely to leave soon due to low job satisfaction and should be offered a raise? Who has high potential and should be put on a fast track career path? Who doesn’t pull their weight and should be warned in their next performance review that their performance puts their job at risk? The systems typically collect data from employees that purportedly allows to quantify anlabod assess them against a set of performance criteria. For some, these systems present a chance to improve both employer and employee satisfaction; for others, it is a further step towards a dystopian society of surveillance and control. In our project, we will investigate the functionalities of the ADM systems in use. What functionalities are on offer, what information is provided to employers and employees about these functionalities, how are staff members informed that their data is collected to measure their performance, and what rights do or should they have to influence or object to the use of the system?
The introduction of mechanisms that automatically evaluate HR data – supporting procedures from recruitment to performance review – has led to imperceptible and lasting changes in the work environment. Agreed written and unwritten rules relating to employment criteria, working hours and workloads have changed, but so too have the expectations of employees and employers in the labour market.
Under the heading of ‘scientific management’, ADM systems are used to optimise production planning and controlling. These systems need data, which in turn makes the quantification of all work processes a prerequisite for their deployment. There are indications from other areas that algorithms can entrench existing biases, for example sex-based or other forms of discrimination. Their deployment could thus hinder the creation of a diverse workforce, an aim that is both desired by business and society. If developed carefully and excluding those biases, these technologies could further this goal by making clear, logic, and objective decisions, improving on those of humans whose decisions are always influenced by subconscious or even conscious prejudices.
- Those who have access to data and the results of its analysis can change the balance of power significantly and in both directions. Is the data produced solely for managers or can everyone view the evaluations and understand or even verify the results?
- Who has access to and control over the collected data? The company, the business that provides the ADM system, the workers’ council? Who should have it? What does meaningful regulation of such systems look like?
- Does the use of ADM in operational control improve or reduce the fairness of performance reviews for employees?
- How does the use of ADM influence the autonomy of employees, their labour rights, and specifically in the German context the aspect of workers’ representation (“Mitbestimmung”) in operational decision making?
We will gather information about the available systems and their use in companies to robustly evaluate their impact on employee autonomy and workers’ rights. We will attempt to identify and explain how the systems work by building model systems with synthetic employee data, and analyse their implications from legal and ethical perspectives. The analysis will allow us to identify potential gaps in regulation arising from both their use and the results. We will also identify ways in which workers’ councils and the wider public should be informed about these systems.
Project period: 1 January 2018 – 30 April 2020
Project lead: Matthias Spielkamp
funded by Hans Boeckler Stiftung