Paper Number

ECIS2026-2630

Paper Type

CRP

Abstract

Although conversational artificial intelligence systems are becoming more prevalent, interactions often suffer from misunderstandings, misalignments, and overreliance. We interpret these issues as a breakdown of ‘common ground’: the shared knowledge, beliefs and assumptions that facilitate successful communication. This paper takes a design science research approach to develop theory-based design patterns to help designers build common ground in human-AI interactions. A systematic literature review was conducted to synthesize five common ground mechanisms. Based on this, we derived five preliminary design patterns that structure recurring contexts, problems, forces, solutions, and consequences. To increase confidence in the maturity and usability of the design patterns, an expert evaluation was conducted. Based on the results, we identified four main changes, which informed the systematic refinement of the design patterns. The resulting pattern set provides intermediate design knowledge to support information systems designers in translating specialized insights on common ground into concrete features for human-GAI interactions.

Share

COinS
 
Jun 14th, 12:00 AM

Towards Design Pattern For Common Ground In Human-Artificial Intelligence Interaction

Although conversational artificial intelligence systems are becoming more prevalent, interactions often suffer from misunderstandings, misalignments, and overreliance. We interpret these issues as a breakdown of ‘common ground’: the shared knowledge, beliefs and assumptions that facilitate successful communication. This paper takes a design science research approach to develop theory-based design patterns to help designers build common ground in human-AI interactions. A systematic literature review was conducted to synthesize five common ground mechanisms. Based on this, we derived five preliminary design patterns that structure recurring contexts, problems, forces, solutions, and consequences. To increase confidence in the maturity and usability of the design patterns, an expert evaluation was conducted. Based on the results, we identified four main changes, which informed the systematic refinement of the design patterns. The resulting pattern set provides intermediate design knowledge to support information systems designers in translating specialized insights on common ground into concrete features for human-GAI interactions.