Keywords

Generative AI, Innovation Diffusion, Case study, Beliefs, Risks, Controversies

Abstract

The reality of how generative artificial intelligences (GenAIs) spread and are used in business is largely unknown. This research aims to describe this in the context of a large technology company, exploring and questioning the ways in which diffusion is accelerated or, on the contrary, the reasons for resistance due to divergent beliefs. Based on diffusionist approaches to innovation and an original questionnaire administered to 1,665 employees, we propose a typology of user profiles. We show that the spread of GenAI is not simply a matter of percolating GenAI systems selected by strategists and spreading them from peer to peer in experiments organized by top management (so-called sandbox experiments). We partition the population into pure experimenters, early and natural adopters, strong early adopters and non-users. Overall, across the different profiles of users and non-users, the initial level of education seems to play an important role in commitment to experimentation, but also in non-use when the level of qualification is low. Those who speak up very often (spreaders) find that GenAI should be generalized are found more often among strong early adopters, while inhibitors are more likely to be found among pure experimenters. Spreaders and inhibitors coexist in professions, creating a fertile ground for controversy. This paper enables us to analyze the diffusion and specific features of GenAI innovation within a large seemingly tech-savvy company. It highlights the need for close support in addressing perceptions and competencies if the goal is to scale up usage within the company.

Download

Share

COinS