Delegating Decisions and Blame to Humans and Machines
13 May 2024
14:30 -16:00
Times are shown in local time.
Open to: All
Room W2.01 (Cambridge Judge Business School)
Trumpington St
Cambridge
CB2 1AG
United Kingdom
Join our Operations and Technology Management seminar
Speaker: Professor Mirko Kremer, Frankfurt School of Finance and Management
About the seminar topic
Despite fast technological advancements and significant investments of financial and managerial resources, the integration of artificial intelligence (AI) technologies into business processes often yields disappointing results. While sometimes the technology itself simply is not sufficiently mature to approach or surpass the performance of existing low tech processes, failed implementation efforts often point to organisational and behavioural issues.
Our study empirically tests a simple theory of algorithmic over adherence, built on anecdotal evidence that is consistent with managers who pressure decision makers to follow algorithmic recommendations while reserving the option to levy blame for bad decision outcomes nevertheless. Such behaviour, besides seemingly inconsistent with fundamental principles of fairness, then expands the purpose of relying on algorithmic advice from improving decisions to sheltering from blame.
To rigorously study the mechanisms underlying such seemingly mis calibrated behaviour, in the context of a simple diagnostic decision task, we first propose a simple informal contracting model with boundedly rational agents that care for distributional and procedural fairness. We then turn to controlled laboratory conditions that allow for a systematic and controlled variation of institutional context, information structure, and available channels for blame. Our experimental data provide support for the predictions of our model analysis: Managers blame decision makers for overwriting algorithmic recommendations, even though they know that such overwrites tend to be in their own best interest. As a result, decision makers under rely on their own task-relevant knowledge, thus limiting the potential value that organisations can reap from implementing AI.
Speaker bio
Mirko Kremer is Professor of Supply Chain Management at the Frankfurt School of Finance and Management since July 2014. He holds a PhD from the University of Mannheim (2008). Before joining Frankfurt School, he was a faculty member at the Pennsylvania State University. His research is on the impact of managerial and customer (mis)behaviour on the performance and design of Supply Chains, with an emphasis on micro-behavioural foundations of inventory management, sales forecasting, and service systems. His research was published in Management Science, Manufacturing and Service Operations Management, Operations Research and Production and Operations Management. He serves as a department editor for the Production and Operations Management Journal, and associate editor for Operations Research.
At Frankfurt School, Mirko teaches operations management, supply chain strategy and quantitative decision modelling. He has previously taught similar courses at INSEAD, London Business School, and the Kellogg School of Management.
Register
No registration required. If you have any questions about this seminar, please email Nigel Low.