Hi, I'm Elisa!
I do research in Natural Language Processing (NLP) and Computational Social Science (CSS), with a particular interest in the interplay between (generative) AI and society. My research examines how the widespread adoption of AI is reshaping social inequalities, creating new forms of disadvantage while also amplifying existing ones. I study how AI technologies serve, or fail to serve, different social groups, depending on their needs and sociolects. My goal is to contribute to the development of more equitable and inclusive AI technologies.
I believe that socioeconomic status (SES) is the strongest driver of inequality in society and there are evidences of how it is now also shaping access to AI, patterns of AI engagement, and the benefits people derive from these technologies. For this reason, I advocate for a greater attention to SES in the development of new technologies.
Before focusing on AI fairness, I worked on Information Extraction. My PhD theses was on cross-domain Relation Extraction: building systems that identify semantic relations between entities and generalize robustly across domains and unseen label spaces.
I am affiliated with the MilaNLP group at Bocconi University, working with Prof. Dirk Hovy. I am also affiliated with the NLP North group at the IT University of Copenhagen and the Pioneer Center for Artificial Intelligence, Denmark.
I completed my PhD at the IT University of Copenhagen, also affiliated with the MaiNLP lab at LMU Munich, supervised by Prof. Barbara Plank and Dr. Rob van der Goot.