Effects of Algorithmic Decision-Making and Interpretability on Human Behavior: Experiments using Crowdsourcing

authored by: Avishek Anand, Kilian Bizer, Alexander Erlei, Ujwal Gadiraju, Christian Heinze, Lukas Meub, Wolfgang Nejdl, Björn Steinrötter
Abstract: Today algorithmic decision-making (ADM) is prevalent in several fields including medicine, the criminal justice system, financial markets etc. On the one hand, this is testament to the ever improving performance and capabilities of complex machine learning models. On the other hand, the increased complexity has resulted in a lack of transparency and interpretability which has led to critical decision-making models being deployed as functional black boxes. There is a general consensus that being able to explain the actions of such systems will help to address legal issues like transparency (ex ante) and compliance requirements (interim) as well as liability (ex post). Moreover it may build trust, expose biases and in turn lead to improved models. This has most recently led to research on extracting post-hoc explanations from black box classifiers and sequence generators in tasks like image captioning, text classification and machine translation. However, there is no work yet that has investigated and revealed the impact of model explanations on the nature of human decision-making. We undertake a large scale study using crowd-sourcing as a means to measure how interpretability affects human-decision making using well understood principles of behavioral economics. To our knowledge this is the first of its kind of an inter-disciplinary study involving interpretability in ADM models.
Organisation(s): L3S Research Centre
Institute of Legal Informatics
External Organisation(s): University of Göttingen
Type: Conference contribution
Publication date: 2018
Publication status: Published
Peer reviewed: Yes
ASJC Scopus subject areas: Computer Science(all)
Sustainable Development Goals: SDG 3 - Good Health and Well-being, SDG 16 - Peace, Justice and Strong Institutions
Electronic version(s): http://ceur-ws.org/Vol-2173/paper5.pdf (Access: Open)