par Ameye, Nicolas 
;Bughin, Jacques ;van Zeebroeck, Nicolas
Référence Technovation, 127, 102846
Publication Publié, 2023-08-25
;Bughin, Jacques ;van Zeebroeck, Nicolas Référence Technovation, 127, 102846
Publication Publié, 2023-08-25
Article révisé par les pairs
| Résumé : | In its recent form, Artificial intelligence (AI) is an ensemble of technologies, which can be defined as machine-based systems for effective enterprise automation and influential decisions”. If businesses that use AI can potentially reap a competitive advantage, the optimal exploitation of such a complex ensemble of technologies remains uncertain as well as requires to have competitive access to complements such as data or new skills. Existing models of organizational use of technologies often ignore either the dynamics of competitive interactions (which can lead to pre-emption or epidemic diffusion) or the role of uncertainty, or both. In the case of AI, one type of uncertainty is particularly important: uncertainty about the technology's use cases (i.e., what to do with it). This paper proposes to apply a real options perspective to the Technology-Organization-Environment (TOE) adoption framework in order to uncover important patterns in the use of AI among firms. The results are threefold: (1) we find evidence of significant epidemic effects in AI use, (2) uncertainty moderates epidemic effects, and (3) the impact of uncertainty depends on its source: an uncertain AI use case reduces herd behaviours while uncertainty about use case returns still favours them. These results highlight the importance of exploration and collective learning in the diffusion of emerging and complex technologies, especially when companies struggle to identify the most profitable use cases for the technology. |
