Fair RecSys - AI-based recommender systems: fairness, transperency and regional economy

The aim of the proposed study on the technology assessment of artificial intelligence (TA study) is to assess the socio-economic potential of AI-based recommender systems. In particular, opportunities for regional value creation and sustainable development are to be discussed. The results of the TA study provide a basis for software developers and political decision-makers alike. Recommender systems support economic decisions by providing information on selected products and services. This mediation service is all the more useful the more confusing the offer is and the better the selection provided can be tailored to the individual needs and preferences of the information seeker. At the same time, the use of AI-supported recommender systems raises important questions of fairness. What criteria are used to evaluate the actors and offers involved and does the provision of recommendations lead to unintentional discrimination against individual interest groups? As Snoboli et al. (2022) argue, the definition of fairness itself is a multi-stakeholder problem that must be negotiated between the groups involved. The proposed TA study takes up this problem and explains it on the basis of a concrete AI development. The results of the proposed study will be incorporated into the development of recommendations for AI-based recommender systems. These recommendations will be communicated to software developers in the form of design guidelines and made available to policy makers in the form of policy briefs.