Design Science Research in an Era of Generative AI—Challenges and Theoretical Guidelines

Information systems (IS) research is increasingly exploring the potential of generative artificial intelligence (GenAI), such as large language models (LLMs). For design science research (DSR), such technologies foster entirely new vistas for the design of IT artifacts that make use of their generative capabilities, but also influence DSR methodology. This shift is much more profound than it has been discussed so far. To identify existing implications of GenAI for design-oriented research in IS, we report results from an integrative literature review of recent DSR publications in leading IS outlets. Thereby, we synthesize five major theoretical challenges that arise when using GenAI in DSR projects: (1) an obscure composition of the artifact, (2) an opaque contextualization of the LLM, (3) a fragile internal consistency of the artifact, (4) a rapid erosion of prescriptive knowledge, and (5) missing methodological guidance. We investigate these challenges and conceptualize a set of three guidelines that inform DSR in the rising era of GenAI. These guidelines support researchers in designing and justifying GenAI-related DSR processes and in precisely articulating the theoretical grounding of their design decisions and evaluation strategies.