Abstract
A fundamental shift in the way society operates is approaching due to the prevalent adoption of self-learning technology like Artificial Intelligence (AI). Defined by the well-pronounced incongruence between their initial purpose and the values delivered to their multifaceted users, AI-powered systems are already being deployed in crucial social institutions such as hospitals, banks and courtrooms. To solve this tension, we first identify design practices that are suitable for the context of AI. Then we introduce a framework of three logical inferences that could aid designers to deliberately and continuously identify and address unintended values AI-powered solutions produce. The paper is concluded by three directions for future research.
Keywords
artificial intelligence; unintended values; purpose
DOI
https://doi.org/10.21606/drs.2020.222
Citation
Stoimenova, N., and Kleinsmann, M. (2020) Identifying and addressing unintended values when designing (with) Artificial Intelligence, in Boess, S., Cheung, M. and Cain, R. (eds.), Synergy - DRS International Conference 2020, 11-14 August, Held online. https://doi.org/10.21606/drs.2020.222
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License
Identifying and addressing unintended values when designing (with) Artificial Intelligence
A fundamental shift in the way society operates is approaching due to the prevalent adoption of self-learning technology like Artificial Intelligence (AI). Defined by the well-pronounced incongruence between their initial purpose and the values delivered to their multifaceted users, AI-powered systems are already being deployed in crucial social institutions such as hospitals, banks and courtrooms. To solve this tension, we first identify design practices that are suitable for the context of AI. Then we introduce a framework of three logical inferences that could aid designers to deliberately and continuously identify and address unintended values AI-powered solutions produce. The paper is concluded by three directions for future research.