A Typological Approach to Analyzing AI Integration in Public Administration

In the face of the advent of artificial intelligence, negative positions arise, pointing to the reduction of jobs, while positive stances focus on the augmentation of the human being and the emergence of new jobs (Julián-Vejar and Velázquez, 2024). However, these approaches tend to focus on the quantitative dimension of the problem, overlooking the qualitative aspects inherent to the world of work (De Stefano, 2019).

The introduction of AI is a reality today, and like any other technique, it must be understood not as an isolated tool, nor as inherently negative or neutral, but as a capacity to build social relationships (Latour, 2018). Based on this line of analysis, this presentation is based on fieldwork conducted within a public organization in France. Although the designers in this entity view the introduction of AI as a mere tool, the reality is that it is a project of greater scope, with implications for the organization and work culture. Including workers in these projects is crucial for a successful and iterative introduction, but public administration has inherent limitations.

For this reason, a typology of analysis is proposed, consisting of four axes to examine the introduction of an AI project in a public institution: “constitution of the project,” “construction of worker acceptance,” “organizational scope,” and “inclusion of workers.” This typology can reveal a vertical project with limited worker inclusion and acceptance focused on their professions, resulting in organizational disruption. However, it can also illustrate a co-construction approach that fosters an understanding of workers' “real activity,” leading to a technological introduction that minimizes disruptions. Based on a study of the various stakeholders involved in an AI project within a French public institution, this typology seeks to improve our understanding of innovation production and the implementation of technological solutions in the public sector.