The Algorithmic Society Chair of MIAI will take a triple challenge to understand this turn : an empirical challenge by studying AI in its immediate social settings, an experimental challenge by proposing innovative cross-disciplinary research on biases in AI and a cultural challenge by fostering a new AI literacy among social science students.
The Chair of MIAI will particularly focus on the impact of the algorithmic turn on social worlds previously governed by institutional and organisational rules (such as the labor market and the public sphere) or by professional knowledge (such as hospitals, courts, newsrooms or city planning offices). It will promote fieldwork research based on well-established social science methods (interviews, observation, text mining and career analysis) and experimentations to avoid the rush to theory and hasty conclusions that AI sometimes provokes. Thus the Chair will help understand the algorithmic turn as a process deeply rooted in society more than as an exogenous revolution.
The Chair will research the social contexts and conditions of AI deployment in different social settings where AI interacts with professional knowledge and organisational constraints in the following fields that chair participants know particularly well : medical treatment in hospitals, judicial decision making in courts, urban planning and newsroom activity. It will also study the perception of AI in the general public, a topic that is key to the future developments of AI and still a blind spot in France. For that purpose large scale text mining research based on the exploration of legacy and social media corpora will be launched. The Chair will also focus on how algorithms permeate daily social practices by studying those who contribute to the elaboration of these algorithms, their trajectories, backgrounds, skills and representations of society.
In addition, the Chair will set up an Open AI Lab fostering research based on the collaboration of social and computer scientists to contribute to a better understanding of the functioning of algorithms and a critical assessment of the effects of AI in society. In collaboration with partners such as AFP and Le Monde, a specific focus will be put on the production of biases and the reduction of viewpoints diversity in the public sphere due to the functioning of algorithmically spread news and comments.
The Chair will finally contribute to the debate on AI in society by organising courses and seminars aiming at developing shared skills and common concerns about AI among social science and computing science students at Univ. Grenoble Alpes. The algorithmic literacy of social science students is still low, as is generally the social science literacy of computer science students. Producing common grounds and research interests that will gather these different students is key to the success of the Chair. This initiative will notably be launched at Univ. Grenoble Alpes with Sciences Po and CitizenCampus.