A team of authors around L3S member Prof. Dr. Avishek Anand won the Best Student Paper Award at the 8th AAAI Conference on Human Computation and Crowdsourcing (HCOMP 2020), which was held virtually on October 25–29th at The Netherlands Institute for Sound and Vision in Hilversum, The Netherlands.
This work studied how the introduction of a decision-support system (DSS) affects human-human interactions. The authors showed that introducing a black-box machine learning model increases the efficacy of human cooperation and overall market outcomes by studying these interactions in the context of a bargaining scenario. They interestingly find that users do not accept the system’s advice always and thereby do not exploit its full potential. Furthermore, economic gains are exclusively driven by improved user decisions. In contrast, people who do not have a DSS themselves become more demanding upon learning that their counterpart has the option to use one. They also find that machine learning models with explanations increases system usage but does not lead to higher system trust or better market outcomes. However, perceptions of unfairness from those not having a system available are mitigated by increased interpretability.
Link to the paper: https://ojs.aaai.org/index.php/HCOMP/article/view/7462