Abstract
Machine Learning (ML) is increasingly becoming a crucial asset across diverse industries. However, designers lack human-centered processes to envision and develop innovative solutions enabled by ML. By engaging in a Research-through-Design activity, we outline a new design process to generate human-centered adaptive systems enabled by data and ML. We describe and discuss the possibilities and limits of designing with ML, the need to concurrently address user experience and ML aspects, and the implications of their mutual influence. We argue that designers can envision and design human-centered ML-enabled systems if they acquire fundamental ML knowledge, although certain tasks necessitate close collaboration with ML experts. We discuss how uncertainty and risk of failure characterize the outlined process and may limit its applicability. The proposed process serves as a foundational framework for future research in human-centered design innovation through data and ML.
Acknowledgments
The authors thank prof. Jean-Bernard Martens for his help with training ML models in the ILLMO environment with our dataset, and for his insights on the ML models performance and results.
Disclosure statement
No potential conflict of interest was reported by the author(s).
Notes
1 This article presents an expanded, reworked version of a conference paper originally published in 14th International Conference of the European Academy of Design, Safe Harbours for Design Research. See Colombo and Costa Citation2021 in the References list for full details.
Additional information
Notes on contributors
Sara Colombo
Sara Colombo, PhD, Assistant Professor at the Faculty of Industrial Design Engineering, TU Delft. She holds a MS in Design & Engineering and a PhD in Design from Politecnico di Milano. Before joining TU Delft, she worked as an Assistant Professor at TU Eindhoven, a Research Scientist at Northeastern University, and a senior Research Associate and Lecturer at MIT. In recent years, she has extensively explored the intersection of AI and design, investigating AI applications for social good across diverse domains, and developing new knowledge and approaches for designing with data and AI. Her current work explores innovative methods and tools for designing human-centered and responsible AI.
Camilla Costa
Camilla Costa, Industrial and UX Designer. She received a Bachelor’s degree in Industrial Design and a Master’s degree in Design & Engineering from Politecnico di Milano. She is interested in how design can foster innovation and generate positive impacts on people’s lives. Her current focus is design diagnostics solutions taking into consideration user needs and sustainability factors.