UdZ 01.21

SPECTRUM – APPLIED RESEARCH 50 / UdZ 2-2021 There were indeed serious supply chain problems in the first phase of the pandemic, so we’re talking about April, May roughly. And that has to do on the one hand with the loss of suppliers in China and then a few weeks later with the loss of suppliers, especially in northern Italy. But that didn’t take long, and already in May and then into the summer, you can talk about a complete normalization of the supplier structures. ” Gabriel Felbermayr, President of the Kiel Institute for the World Economy “ T he past has shown that individual crises such as pandemics, natural disasters or major political changes always burden several social, political and economic areas due to their complexity 1 . The consequences permeate all social structures and thus affect every individual. Omnipresent is the Corona pandemic, which exemplarily reveals how well-timed and global supply chains are negatively affected. One of the reasons for this lies in the popular principles of lean management. Processes are designed to be efficient and wasteful. 2 As a result, they serve as exemplary drivers for clocked just-in-time production. However, the basis for this is a set of stable processes. The disruption of individual processes is assumed to be fundamentally surmountable. In “normal” everyday business, potential disruptions can be repeatedly handled without major losses. In crises, however, disruptive disturbances occur that can no longer be managed according to the well- known principle of “Actio - Reactio”. The unpredictability of the causes, duration and severity of a crisis make preparation an ambiguous question of complexity. Predictive and data-based measures can be used to manage the resulting obstacles. 3 Digital transformation is already generating big data for this purpose, including in healthcare. This potential can be processed with increasingly powerful computing machines and used for intelligent detection of crises 4 . Nevertheless, this is sensitive data that is correctly not freely accessible. However, there is a further problem in the concatenation of anonymized data in the practical implementation of crisis management: In particular, the lack of a cross-domain ecosystem for crisis management as well as the lack of data sovereignty and the corresponding data protection, the few resources in small and medium-sized enterprises as well as the lack of integration of proprietary data and closed systems form the core challenges. 5 The crisis management platform to be developed in the ‘PAIRS’ project should help to successfully overcome these hurdles in the future. Ensured by a cross-domain and federated approach, it will serve to identify and anticipate crisis scenarios on the basis of a hybrid AI concept (the combination of human and artificial intelligence) in an iterative learning process. The aim is to secure the availability of essential resources and capabilities of companies and their networks and to sustainably strengthen their marketability. In particular, the interactions between economic and political actors will be considered, taking into account data sovereignty and data security. In the realization of this AI lighthouse project, more than 25 partners as well as the FIR are already involved with expertise from production and information management. The platform reflects the following cycle: the risk of individual crisis scenarios, i.e. concrete manifestations of the crisis event and the general reactions, are specifically assessed in relation to the trigger of the crisis or to the respective trouble spot. From this risk assessment and a self-expanding pool of measures, individual response measures to possible 1 Petersen and Bluth 2020, p. 9 2 Görg et al. 2020, p. 7 3 Stich et al. 2021 , p. 4 Schäfer 2021 5 Stich et al. 2021, p. 35 et seq.