Guest: Prof. dr. Richard Rogers (profile)
Affiliation: Media Studies and Digital Methods Initiative, Humanities Labs, University of Amsterdam
We invite you to an event organized by the Department of Political Science, Faculty of Arts, Pavol Jozef Šafárik University in Košice.
The lecture will present a critical discussion on the use of Large Language Models (LLMs) in research.
It will focus on prompting and classification — how models can be involved in the process of data annotation and categorization, the methodological challenges this entails, and the risks of uncritical use of outputs. The lecture will also address the issue of sensitivity to “guardrails” — the models’ ability to respond to restrictions, rules, and safety mechanisms.
Opportunities for employing auditing to assess the reliability and transparency of LLMs as well as exploring the boundaries of generating synthetic data will be discussed.
The event is held within the project Audit of the level of internationalization and implementation of internationalization projects of higher education institutions and research institutions.
The lecture and discussion will be held in English.
PROGRAM:
9:15 – Opening and introductions
9:30–10:15 – Research-with-AI critique: Prompting & Classifying
10:15–10:45 – Coffee break
10:45–12:00 – Research-with-AI critique: Auditing & Guardrail Sensitivity
