preprint OISI26 t-10 inst_AI2 - Communicative Infrastructures for Organisational AI: Why Generative Systems Require Accountability Architectures
Summary
No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.
Summarise with VIRAT Free to try with your Science Catalogs account.Abstract
Large language models are rapidly entering organisational communication workflows, where they are used to draft emails, recommendations, reports, and operational messages. This development signals a shift in the design problem of information systems from managing information to sustaining accountable communication. Although generative systems produce fluent text, fluency alone does not establish the institutional standing of a communicative act. In organisational settings, expressions such as proposals, approvals, and instructions acquire meaning only when issued by identifiable actors operating in recognised roles within communicative processes that support accountability. When AI-generated text circulates without explicit institutional framing, organisations may struggle to reconstruct how decisions emerged and who assumed responsibility for them. This paper argues that the effective integration of generative AI requires attention to the communicative infrastructures through which organisational acts become accountable. By bringing into dialogue research on generative AI, the language–action tradition, and socio-technical studies of digital infrastructures, the paper introduces the concept of communicative infrastructures and proposes accountability architectures as a minimal design response to AI-mediated organisational communication. The contribution is conceptual and design-oriented, offering a new perspective on how information systems should be theorised and designed in the age of generative language technologies.
Fields of Study & Subjects
Information Systems Theories and Implementation
Ethics and Social Impacts of AI
Cybernetics and Technology in Society
Publication details
- Journal / source
- AIS Electronic Library (AISeL)
- Publication date
- 29 March 2026
- Language
- EN