In a context of global competition, governments encourage and support the development of AI, including in the public sector. The promises made are varied: savings in costs, the ability to consider variables that human understanding alone cannot see, automating user requests processes and their response, automating decision-making. Besides, the challenge is also to contribute to the well-being of citizens, while guaranteeing their fundamental rights. Public action is not only the action of governments to pursue this agenda. This collective action can thus be observed as the self-government of common resources by a wide range of actors. How does AI impact the terms of this action? Therefore, we propose to confront Elinor Ostrom's theoretical framework of commons - in particular the criteria of institutional performance – to the analysis of EU Joint Research Centre database, presenting characteristics of 686 AI cases in the EU public sector.

Abstract Only