Collaborative Rationality and Evaluation in Data-Driven Decision- Making. From theory to a model and design principles
Thesis event information
Date and time of the thesis defence
Place of the thesis defence
L10, Linnanmaa campus
Topic of the dissertation
Collaborative Rationality and Evaluation in Data-Driven Decision- Making. From theory to a model and design principles
Doctoral candidate
Master of Science Nada Sanad
Faculty and unit
University of Oulu Graduate School, Faculty of Information Technology and Electrical Engineering, Empirical Software Engineering in Software, Systems and Services (M3S)
Subject of study
Information Processing Science
Opponent
Professor Kalle Lyytinen, Case Western Reserve University
Custos
Professor Tero Päivärinta, University of Oulu
Human-machine collaboration for enhancing data-driven decision-making
The nature of decision-making has changed due to advancements in artificial intelligence (AI), machine learning, data science, and analytics. This research proposes new methods to enhance data-driven decision-making (DDDM) by augmenting human and machine intelligence and capabilities, leading to better collaboration, outcomes, and learning.
A novel theory is presented which claims that DDDM has evolved to include five main elements. These elements are the human decision-maker, decision-making process, decision and its outcomes, massive amounts and types of data, and augmented capabilities of machines. Only by considering each of these elements individually, and their relationship as a whole, can DDDM be successfully implemented in organizations. However, organizations and developers still struggle to design systems that support the multi-faceted nature of DDDM.
By adopting a design science research methodology, a conceptual model is proposed to help organizations plan and design systems that incorporate the DDDM elements, enhance collaboration between humans and machines, and support the evaluation of decisions after they are made. This ex-post evaluation is crucial as it allows for different types of continuous learning and improvement, as well as integrating feedback into future decisions. Finally, a set of design principles is recommended to help developers build technologies which implement the model in practice.
The model and design principles were practically demonstrated and evaluated in the case of AI-enabled menu design at Antell restaurants to enhance DDDM, decision evaluation, and human-machine collaboration and learning. They are robust and adaptable enough to be applied to various use cases and decision scenarios, and provide actionable tools and statements to support implementation and system development in organizations. By defining and modelling the DDDM elements, system developers and decision stakeholders can have a unified viewpoint, leading to increased transparency, accountability, and AI adoption. Human and machine learning is enhanced through evaluation, and future decisions are more informed.
From a theoretical perspective, this research enriches our understanding of human-machine collaboration in DDDM. It provides a roadmap for future research in designing systems that better integrate human and machine intelligence and capabilities for enhanced decision-making.
A novel theory is presented which claims that DDDM has evolved to include five main elements. These elements are the human decision-maker, decision-making process, decision and its outcomes, massive amounts and types of data, and augmented capabilities of machines. Only by considering each of these elements individually, and their relationship as a whole, can DDDM be successfully implemented in organizations. However, organizations and developers still struggle to design systems that support the multi-faceted nature of DDDM.
By adopting a design science research methodology, a conceptual model is proposed to help organizations plan and design systems that incorporate the DDDM elements, enhance collaboration between humans and machines, and support the evaluation of decisions after they are made. This ex-post evaluation is crucial as it allows for different types of continuous learning and improvement, as well as integrating feedback into future decisions. Finally, a set of design principles is recommended to help developers build technologies which implement the model in practice.
The model and design principles were practically demonstrated and evaluated in the case of AI-enabled menu design at Antell restaurants to enhance DDDM, decision evaluation, and human-machine collaboration and learning. They are robust and adaptable enough to be applied to various use cases and decision scenarios, and provide actionable tools and statements to support implementation and system development in organizations. By defining and modelling the DDDM elements, system developers and decision stakeholders can have a unified viewpoint, leading to increased transparency, accountability, and AI adoption. Human and machine learning is enhanced through evaluation, and future decisions are more informed.
From a theoretical perspective, this research enriches our understanding of human-machine collaboration in DDDM. It provides a roadmap for future research in designing systems that better integrate human and machine intelligence and capabilities for enhanced decision-making.
Last updated: 3.5.2024