Solving complex problems was named a new challenge for research on human-AI collaboration. In our study we focus on a particular means to solving complex problems: design science research (DSR). We investigate whether AI, more specifically, generative language models (GLM), can support an individual in conceptualizing DSR studies by making helpful suggestions. To do so we use extracts of a published DSR study and have GPT-3, a GLM provide suggestions for aspects of this study. These suggestions are then evaluated in a survey (n=33) regarding their helpfulness. Results show that GLM suggestions are perceived to be helpful, with some variation depending on expertise. Reported interest in using such a tool in the future was high. Describing how GLM can offer helpful suggestions we contribute toward a DSR tool support ecosystem and, more generally, towards knowledge on how humans and (generative) AI systems can team up to solve complex problems.
Memmert, Lucas; Cvetkovic, Izabel; and Bittner, Eva, "HUMAN-AI COLLABORATION IN CONCEPTUALIZING DESIGN SCIENCE RESEARCH STUDIES: PERCEIVED HELPFULNESS OF GENERATIVE LANGUAGE MODEL'S SUGGESTIONS" (2023). ECIS 2023 Research Papers. 405.