SUMMARY

This workshop on Bias in Generative AI focuses specifically on text-generating technologies called Large Language Models (LLMs), such as ChatGPT.  

Through an innovative combination of participatory methods, developed by University of Edinburgh PhD student Lara Dal Molin, the workshop will explore concrete instances of bias in an open-source LLM.  

Firstly, Lara will introduce the basic functionalities of LLMs and the emerging issue of bias within these. Then, the session will be articulated into three activities - Prompting, Scoring and Designing Stories - that participants will carry out in groups. These activities will be facilitated by University of Edinburgh and Edinburgh Futures Institute students Carla Garcia, Savannah Thomson and Antara Sutradhar. In Prompting, participants will be invited to discuss relevant dimensions of their identities and guided through a prompt engineering exercise. In Scoring, participants will observe model outputs and reflect on their detectable extent of bias. In Designing Stories, participants will analyse emerging themes in the model's responses and envision values-oriented alternatives.