
Salim Hafid
Présentation
(last updated: July 2025)
I am currently a postdoctoral researcher at the médialab at Sciences Po, where my research aims to understand the impact of AI on democratic processes such as voting or deliberation. I am a member of the “Democratic Commons” research project, an interdisciplinary project which aims at measuring the impact of AI on democracy. The project involves researchers from political theory, ethics, sociology, and computer science to answer the following question: Do AI systems exhibit democratic biases (e.g political orientation) that could influence the users they interact with? The project is funded by the BPI and conducted by a research consortium including the médialab, CEVIPOF, ISIR, and Make.org.
Before joining SciencesPo, I was trained as an engineer at the National Institute of Applied Sciences (INSA de Lyon), and I hold a PhD in Computer Science from the University of Montpellier, completed as part of the “AI4Sci” european project in collaboration with the Leibniz Institute for Social Sciences (GESIS) in Germany. My doctoral research focused on the study of scientific discourse on social media and led to novel methods which aim at improving the quality of scientific debates online and facilitating fact-checking.
More broadly, my research interests include AI for democracy, computational social sciences, and the analysis and modeling of online discourse. I am member of the organizing committee for the international CheckThat! workshop, and a recipient of the “NSF/SIGWEB Student Award” for my work presented at the international CIKM2022 conference.
Publications
Publications
An In-depth Analysis of the Linguistic Characteristics of Science Claims on the Web and their Impact on Fact-checkingACM Transactions on the Web, 2025, ⟨10.1145/3746170⟩
Article dans une revue
hal-05163522
v1
|
Disambiguation of Implicit Scientific References on XHT 2025 - 36th ACM Conference on HyperText and Social Media, Sep 2025, Chicago, United States
Communication dans un congrès
hal-05163668
v1
|
|
|
Cite-worthiness Detection on Social Media: A Preliminary StudyNSLP 2024 - 1st International Workshop on Natural Scientific Language Processing and Research Knowledge Graphs, May 2024, Hersonissos, Greece. pp.19-30, ⟨10.1007/978-3-031-65794-8_2⟩
Communication dans un congrès
hal-04801702
v1
|
Investigating Characteristics, Biases and Evolution of Fact-Checked Claims on the WebHT 2024 - 35th ACM Conference on Hypertext and Social Media, Sep 2024, Poznan Poland, Poland. pp.246-258, ⟨10.1145/3648188.3675135⟩
Communication dans un congrès
hal-04801517
v1
|
|
|
SciTweets - A Dataset and Annotation Framework for Detecting Scientific Online DiscourseCIKM 2022 - 31st ACM International Conference on Information and Knowledge Management, Oct 2022, Atlanta, GA, United States. pp.3988-3992, ⟨10.1145/3511808.3557693⟩
Communication dans un congrès
hal-04479646
v1
|
A Hybrid Approach for Stock Market Prediction Using Financial News and StocktwitsCLEF 2021: Experimental IR Meets Multilinguality, Multimodality, and Interaction, Sep 2021, Virtual conference, France. pp.15-26, ⟨10.1007/978-3-030-85251-1_2⟩
Communication dans un congrès
hal-04020259
v1
|
|
Détection, liage et interprétation d'énoncés scientifiques et leurs contextes à partir de discours en ligneInformatique [cs]. Université de Montpellier, 2024. Français. ⟨NNT : 2024UMONS044⟩
Thèse
tel-04951515
v2
|