Co-Shaping an Ecosystem for Responsible AI: Five Types of Expectation Work in Response to a Technological Frame

• Published in: Information systems frontiers 2023-02, Vol.25 (1), p.103-121
• Main Authors: Minkkinen, Matti, Zimmer, Markus Philipp, Mäntymäki, Matti
• Format: Article
• Language: English
• Subjects: Algorithms; Artificial intelligence; Business and Management; Collaboration; Control; Ecosystems; Ethics; Frames; Information systems; IT in Business; Management of Computing and Information Systems; Operations Research/Decision Theory; Systems Theory; Technology
• ISSN: 1387-3326; 1572-9419
• DOI: 10.1007/s10796-022-10269-2

Governing artificial intelligence (AI) requires cooperation, although the collaboration’s form remains unclear. Technological frames provide a theoretical perspective for understanding how actors interpret a technology and act upon its development, use, and governance. However, we know little about how actors shape technological frames. In this paper, we study the shaping of the technological frame of the European ecosystem for responsible AI (RAI). Through an analysis of EU documents, we identified four expectations that constitute the EU’s technological frame for the RAI ecosystem. Moreover, through interviews with RAI actors, we revealed five types of expectation work responding to this frame: reproducing, translating, and extending (congruent expectation work), and scrutinizing and rooting (incongruent expectation work). Furthermore, we conceptualize expectation work as actors’ purposive actions in creating and negotiating expectations. Our study contributes to the literature on technological frames, technology-centered ecosystems, and RAI while also elucidating the dimensions and co-shaping of technological frames.