The role of algorithm aversion and human-automation trust in the use and acceptance of Forecast Support Systems

01 October 2020 → 31 December 2022
Research Foundation - Flanders (FWO)
Research disciplines
  • Social sciences
    • Knowledge representation and machine learning
    • Cognitive science and intelligent systems not elsewhere classified
    • Research methods and experimental design
    • Applied economics not elsewhere classified
    • Logistics and supply chain management
algorthm aversion
Project description

Forecasting is a vital skill needed for remaining competitive in today’s business world. Forecast Support Systems (FSS) have now largely automated the forecasting process. Yet, the role of the forecasters themselves cannot be underestimated. Judgment occurs at various stages of the forecasting process, such as in the selection and evaluation of a forecasting model. But just how good are people at working with formal methods? Two distinct but related concepts are of importance here: algorithm aversion and trust in human-automation pairings. Algorithm aversion refers to the fact that people tend to judge algorithms more harshly than they do humans, and are quick to stop using them. Moreover, we distrust algorithms and various other forms of automation. If we want to increase the use of Forecast Support Systems and their trust in them (and thereby increase effectiveness of forecasts), we need a solid understanding of the factors surrounding aversion towards and distrust in them. Only then will we be able to implement the remedial measures.