We undertake research along three main axis:
This includes in particular:
we use multiagent systems to simulate complex socio-economic human systems. In particular we recently worked on the modelization of :
We are contributing to the design and development of:
E-democracy is a form of government that allows everybody to participate in the development of laws. It has numerous benefits since it strengthens the integration of citizens in the political debate. Several on-line platforms exist; most of them propose to represent a debate in the form of a graph, which allows humans to better grasp the arguments and their relations. However, once the arguments are entered in the system, little or no automatic treatment is done by such platforms. Given the development of online consultations, it is clear that in the near future we can expect thousands of arguments on some hot topics, which will make the manual analysis difficult and time-consuming. The goal of this project is to use artificial intelligence, computational argumentation theory and natural language processing in order to detect the most important arguments, estimate the acceptability degrees of arguments and predict the decision that will be taken.
See website
*The AI for Augmented Deliberation collaborative project (PAC) is part of the PostGenAI “ClusterAI” consortium lead by SCAI (Sorbonne Consortium for AI). The project started in January 2025 and is funded for 5 years (ANR-23-IACL-0007). It is jointly coordinated by Nicolas Maudet at LIP6 and Jean-Philippe Cointet at Sciences Po and it involves Nukk.AI and Make.org as industrial partners.
See website
The SCIENTIA project (2022-2026) aims to study the mechanisms that foster the diffusion of AI technologies, the effects of this breakthrough innovation on scientific cultures and practices, and the societal implications of the AI-driven changes to the scientific landscape. The project is led by Floriana Gargiulo (GEMASS), with the involvement of the Centre Internet et Société, LIP6, and Sony CSL as partner organizations.
Within the SMA team, Jean-Gabriel Ganascia (lead), Gauvain Bourgne (co-lead), Evangelia Zve (associated phd student), and Benjamin Icard (associated postdoc) are involved in the thematic monitoring of scientific domains affected by the use of AI and in tracking their evolution and reconfiguration over time.
See website
The THEMIS project aims at producing a general ordinal theory of cooperative interaction situations and power indices for the formulation of a portfolio of social ranking solutions applied to different domains of artificial intelligence, such as decision theory, game theory, computational social choice and multi-agent systems. In particular, the project will focus on the following families of problems: 1) the axiomatic design of novel social ranking solutions accompanied with a road-map of principles guiding users to the most adapted scenario; 2) the impact of the computational difficulties of algorithms for social ranking and their vulnerability to strategic behaviour; 3) the dynamics of coalition formation and the effect of social ranking solutions on the behaviour of individuals to form stable coalition structures; 4) the application of compact preference representation to efficiently compute social ranking solutions and to assess their robustness to changes and to manipulations. Each family of problems from (1) to (4) is studied in a specific work package, and a fifth work package is devoted to the implementation and to the application of our models and solutions to case-studies from voting theory, the analysis of influence measures on social networks, argumentation theory and multi-criteria decision analysis. Specific examples of application will concern the ranking of European countries according to their effective impact over EU public policies; the interpretation of weights for the electoral college of members of the Paris Sciences et Lettre (PSL) University; the comparison of influence, popularity, trends, etc., over Twitter and other social networks; the definition of gradual semantics to rank arguments with regard to their overall strengths in a debate; the representation of positive and negative synergies between criteria in a multi-criteria decision analysis problem.
See Website
How to mitigate efficiently global warming et protect biodiversity? The TerraNeon project aims at designing a decision-aiding tool to help companies and public actors to monitor their impact on the environment efficiently.
See Website
The TRUSTEDNEWS project (2026-2028) aims to develop an integrated tool for analyzing the quality of textual documents and detecting information manipulation in the printed press. TRUSTEDNEWS is led by Paul Égré at IRL Crossing (CNRS), with the involvement of LIP6 and IRISA, as well as partner organizations AIRBUS and OUEST-FRANCE.
The SMA team is involved through Jean-Gabriel Ganascia (co-lead), Benjamin Icard (initiator and architect, postdoc), Lila Sainero (associated research engineer) and Gauvain Bourgne (lead), focusing on developments, experiments and analyses in natural language processing. Particular attention will be given to the temporal tracking of manipulation attempts and to the role played by generative AI. Through its industrial partners, TRUSTEDNEWS aims to test the integrated tool in realistic environments and on new use cases. The aim is to provide military and civilian users with clear, multidimensional indications of the level of trust or distrust that should be placed in the content of a document or a set of online press documents.
See Website
The VIGILANT project (2026-2029) aims to measure, understand, and reduce individuals’ cognitive vulnerability to online information manipulation, in a context of growing cognitive warfare targeting civilian and military populations. The project will develop an operational digital tool assessing individual susceptibility to influence, identifying the narrative triggers exploited in propaganda operations, and delivering personalized countermeasures to strengthen cognitive resilience. VIGILANT is led by Camille Salinesi at the CRI (Université Paris 1 Panthéon-Sorbonne), with the involvement of LIP6 and the industrial partner Everdian, as well as European laboratories (Tartu, Tilburg, Neu-Ulm) as subcontractors for cross-cultural experimentation.
Within the SMA team, Gauvain Bourgne (lead), Benjamin Icard (initiator and architect, associated postdoc), Lila Sainero (research engineer) and Jean-Gabriel Ganascia (co-lead) are involved in developing the digital framework and integrating the linguistic and cognitive analysis modules for the detection of hostile narratives, linking NLP-extracted narrative structures to psychometric vulnerability profiles. Through its industrial partner, VIGILANT aims to validate the tool in operational conditions, providing military and civilian users with reliable measures of susceptibility to influence and adaptive training in informational vigilance.
SimParc (Simulation de gestion participative de parcs) est un projet de recherche visant à explorer des méthodologies participatives de gestion de parcs et espaces protégés pour la conservation de la biodiversité et l’inclusion sociale.
La fondation du projet est un jeu sérieux distribué simulant un conseil de gestion de parc ou d’espace protégé et la dynamique du processus de décision et de négociation parmi ses membres (des représentants d’acteurs sociaux – « stakeholders » –, par exemple : représentant d’ONG environnementale, opérateur de tourisme, représentant de population traditionnelle, syndicat de pêcheurs, etc.). Les différents joueurs jouent le rôle des différents acteurs sociaux correspondant à un scénario donné (parcs nationaux, parcs régionaux, réserves terrestres, marines, extractivistes, de biosphère…) comprenant un ou plusieurs gestionnaires, différents représentants d’acteurs sociaux et des modes de conseil consultatif ou décisionnaire. Les scénarios étudiés sont par exemple : le processus participatif de zonage (détermination des niveaux de conservation) des différents secteurs de l’espace protégé, la politique de gestion (des ressources biologiques, par exemple des poissons) et d’intervention (nettoyage, repeuplement…) de réserves, etc.
Les prototypes informatiques actuels combinent différentes techniques : jeux sérieux (jeux de rôles distribués informatisés), interfaces spécialisées, agents joueurs, agents assistants, aide à la décision, analyse de viabilité, argumentation, aide à la négociation. Les techniques développées sont dans les domaines de : intelligence artificielle, interaction homme machine, aide à la décision et mathématiques appliquées.
The NEGOCLIM project brings together researchers in economics and computer science to study climate negotiations.
See Website