Arrangör

AI Lund

Publicerad på

Beskrivning

Welcome to a seminar where we will find out more about how interactions between large language models (LLMs) and humans shape judgment, decision-making, and the interpretation of information. As LLMs become increasingly embedded in work, research, and everyday decision-making, it becomes important to understand not only how people use them, but also how these systems behave. While LLMs can support analysis and improve tasks such as sentiment assessment, they may also influence human judgments and reproduce biases similar to those found in people. In this seminar, Professor Daniel E. O'Leary will share his research on human-LLM interaction, including experiments showing that people can be biased by LLM responses. The presentation will also examine how LLMs can display anchoring bias, confirmation bias, specificity bias, variation in sentiment estimates, avoidance of extreme judgments, and an ostrich effect. The findings show both the promise and limitations of LLMs: they can outperform human judgments in some tasks, but they also introduce new risks that must be understood when applying them in research, organizations, and decision-making. In the end, the speaker will also examine some implications of these findings for different applications. As part of the LUSEM International Collaboration Initiative, launched by LUSEM’s Vice-Dean Professor Jonas Helgertz, we are pleased to invite you to a seminar in this series. The seminar will be delivered by Daniel E. O’Leary, Professor at the Marshall School of Business, University of Southern California, United States. Professor O’Leary is visiting Blerim Emruli, Senior Lecturer at the Department of Informatics.

Talare

Målgrupp

Taggar