A new data set is now available online in the Qualiservice catalog at https://doi.org/10.1594/PANGAEA.961744.

These are 18 transcripts of qualitative expert interviews from the Hans Böckler Foundation-funded project Algorithmic Prediction and Co-Determination (AVuM), which was jointly conducted by Maximilian Heimstädt (Weizenbaum Institute) and Lukas Daniel Klausner (St. Pölten University of Applied Sciences) from October 2021 to December 2023.

The project investigates the use of algorithmic methods for risk management ("Predictive Risk Intelligence" or PRI) in increasingly complex value creation networks that predict the probability of failure of machines or infrastructure. The central questions were how and with what consequences PRI is already being used by companies for company and inter-company co-determination or how algorithmic prediction systems (similar to PRI or new) can be used by employee representatives to further develop co-determination in times of desolidarization. A secondary object of investigation was also the impact of the Supply Chain Duty of Care Act (LkSG) introduced in Germany on risk management practices.

This topic was investigated through qualitative, semi-structured expert interviews with employees in companies, NGOs or legal interest groups in Germany and Austria. To this end, guided expert interviews were conducted over a period of eight months (June 2023 - February 2023) with thirty representatives of three stakeholder groups (merchants, customers and audiences) relating to PRI in supply chains. Merchants are representatives of software solutions for risk management, the audience group includes interest groups for risk management in supply chains, and customers are (potential) customers of PRI providers who mainly come from the areas of purchasing and supply chain management. The interviews focused on the classification of their own organization in the context of global supply chains and risk management, experiences or assessments of the use of PRI and relationships with other stakeholder groups.

The data collected was evaluated through a reflexive thematic analysis and, due to the thematic breadth of the interviews conducted, offers the potential for further research, in particular on the dynamics between companies, interest groups and risk management service providers and on the effects of algorithmic prediction methods on various aspects of labor relations. The anonymized transcripts are available as a scientific use file for secondary scientific use and for academic teaching.