Explainable and responsible artificial intelligence
Christian Meske, Babak Abedin, Mathias Klier, Fethi Rabhi · Electronic Markets · 2022
Today’s algorithms already reached or even surpassed the task performance of humans in various domains. Especially, Artificial Intelligence (AI) plays a central role for the interaction between organizations and individuals such as their customers, transforming for instance electronic commerce or customer relationship management. However, most AI systems are still “black boxes” that are difficult to comprehend—not only for developers, but also for consumers and decision-makers (Meske et al., 2022 ). With regards to electronic markets, problems such as trying to manage the risk and ensure regulatory compliance of electronic trading systems based on machine learning stem not only from their data-driven nature and technical complexity, but also from their black-box nature, where the “learning” creates non-transparent dependencies between inputs and outputs (Cliff & Treleaven, 2010 ). This raises many challenges such as ensuring data quality issues, managing provenance information needed for transparency as well as organizing metadata when combining data from multiple sources (Rabhi et al., 2020 ). Thus, a responsible and more trustworthy AI is demanded (HLEG-AI, 2019 ; Thiebes et al., 2021 ; Schneider et al., 2022 ).