How can we bring more transparency to automated decision making (ADM) processes that have a significant impact on our societies? And what consequences has the use of ADM systems in human resources management, and therefore for workers’ rights and co-determination (employee representation)?
We are pleased to announce that we can address these issues in the coming months and years with the support of the Bertelsmann Stiftung and the Hans Böckler Foundation. Total funding by both organisations amounts to to a mid-six-digit sum.
AlgorithmWatch (http://algorithmwatch.org) is a non-profit research and advocacy organisation. We monitor the use of ADM systems in fields relevant to society because we are convinced that ADM processes can be used to strengthen inclusion. They should not be used to weaken participation in society.
To achieve this aim we have set out to watch, explain, network and engage. We analyse the effects of ADM processes on human behaviour and point out ethical conflicts. We explain the characteristics and effects of complex ADM processes to a general public and convene experts from different cultures and disciplines focused on the study of these processes and their impact on society. And last but not least we explore ideas and strategies to achieve accountability of these processes, i.e. by developing best practices to be considered during the design process of those systems and by devising norms and standards for the use of ADM systems.
The support we receive from Bertelsmann Stiftung will be used towards the first two pillars of our mission, to watch and explain. We will strive to strengthen transparency by pointing out which ADM systems are used and for which purposes. What do we know about these systems, how are they regulated? And how can we explain their function and their effects to non-experts? It is essential to gather this knowledge and insight to in order to foster an informed public capable of making meaningful decisions about the use of these systems.
Funds by Hans Böckler Foundation are earmarked for our research project Automated human resources management and its effects on co-determination(employee representation). Here, we will examine ADM systems used to evaluate and manage employee performance. These systems are used to predict employees’ performance and behaviour, such as likeliness of attrition. We want to find out which functions the systems employed offer to users (employers) and those affected (employees), what information employees have about the use of the systems and what say they have in their deployment and use.