CAIS researcher Ahrabhi Kathirgamalingam, together with Tariq Choucair (lead and presenter), Laura Vodden, and Axel Bruns, is involved in a paper being presented at AoIR2025 in Brazil. It will be presented in the panel “Humans and Machines in the Loop: Rethinking LLMs for Conflict and Disagreement in Content Analysis of Social Complex Phenomena.”
The paper provides a comprehensive literature review on how LLMs (Large Language Models) are transforming content analysis in the social sciences. It describes four central modes of LLM integration: scalable coders, human-assisted collaborators, autonomous decision-makers, and tools for semantic clustering. The study also highlights ongoing challenges, particularly regarding interpretability, reliability, and epistemic authority when applying LLMs to human-generated texts.
The presentation will take place on Thursday, October 16, from 9:00 to 10:30 AM on-site.
Further information about the panel is available at:
https://www.conftool.org/aoir2025/index.php?page=browseSessions&form_session=547#paperID683