Artificial intelligence (AI) has been a prominent topic of discussion, particularly with the emergence of large language models like ChatGPT, which stirred both awe and concern earlier this year. These concerns range from issues such as plagiarism to existential fears. Nevertheless, AI is increasingly becoming an integral part of human activity especially in decision-making processes.
Within the realm of management and organisational decision-making, AI has raised concerns on two fronts: the underutilisation of AI when it could be beneficial and the overreliance on AI, potentially leading to the loss of unique human knowledge and an aloof mentality towards decision-making. L. Valtonen's dissertation delves into the intricacies of decision-making, not merely focusing on whether AI is over- or underused but examining how its usage influences decision rationales.
"What I discovered is nuance and paradox: AI suggestions and assistance in decision-making are often perceived as more objective and rational. At the same time, however, the rationality and reasons behind these decisions are frequently hidden or incomprehensible to us. How can we confidently assert their increased rationality, especially given the potential biases in AI resulting from historical data?” Valtonen asks.
Valtonen’s research revealed that AI tends to conceal decision rationality in various stages of the decision-making process, yet, in certain instances, it also unveils possibilities for unique and novel rationales. Two levels of rationality concealment were identified: the content of a rationality and the existence of a rationality in the first place. The results suggest that individuals can become involved in their own alienation from rationality in the context of AI decision-making, a reflection that the dissertation also explores.
Valtonen’s contribution to the discourse on AI and decision-making lies in proposing testable claims about the impacts of AI, offering avenues to enhance fairness, accountability, and transparency. The research provides guidance for practitioners to discern and navigate the effects of AI on decision-making, facilitating successful adoption and addressing potential challenges across diverse organisational contexts.
Originally from Kokkola, Valtonen currently works at IdeaSquare, the innovation space at CERN, in Switzerland.
Public defence on Friday 20 October
The doctoral dissertation of MSc (Tech) L. Valtonen titled Rationality in Artificial Intelligence Decision-making will be publicly examined at the Faculty of Management and Business of Tampere University at 12 o’clock on Friday 20 October 2023 in auditorium Pieni Sali 1 of the Festia building (address: Korkeakoulunkatu 8, Tampere). The Opponent will be Emeritus Professor Timo Airaksinen from the University of Helsinki while Professor Miia Martinsuo from Tampere University will act as the Custos.