Examine the inherent biases in the training data of generative image AI, with a particular focus on the stereotypes for terms related to Classical Studies. This project will involve testing programmes such as Midjourney, Stable Diffusion and Dall-E-3.
Department: Classics
Supervised by: Jackie Baines; Edward A S Ross
This project is part of the wider 'ChatGPT: A Conversational Language Study Tool' TLEP project in the Department of Classics which examines the impact and effectiveness of using generative AI for teaching and learning in the field of Classics. With the growing importance of generative AI, it is crucial for students to understand the how these tools work, the ethics involved in their training, and the problems with using them. The aim is to further the understanding of the biases in generative AI, in particular for any output related to topics in Classical Studies. The student involved in this project will be researching case studies using two to three generative AI Image tools. For the first week of the project the student will try out different AI generative Image tools and read about their training data to discern which tools would be most effective for their case studies. Weeks 2 to 4 will be dedicated to their three chosen tools, producing outputs to test their hypotheses about AI bias. During this time they will compare outputs to ancient art and classical themes in popular culture, using appropriate data bases. Weeks 5 and 6 will be dedicated to writing up a report of their findings. During the entire project they will be working closely with Co-Is who will also be investigating biases in generative AI tools for a broader paper on the subject. The student's cases studies will be included in the paper and they will be given co-author credit.
• Engaging In research with scholarship on generative AI training data • Using and testing AI image tools to determine the most effective tools • Rigorous testing of the tools and comparing outputs to popular culture and archaeological material • Writing a report of their findings and refining the text to prepare it for publication
Essential • Able to maintain detailed research notes • Able to work independently • Be familiar with themes in Classical Studies and of classics in popular media Desirable • Experience with project work • Interest in Artificial Intelligence technology
• Building AI competency • Developing research skills with museum data bases • Academic writing skills for publication • Skills for working on collaborative academic projects • Co-authorship on an academic journal article
Edith Morley Building, Department of Classics, University of Reading
20 hours per week (flexible start and end times)
Monday 10 June 2024 - Friday 26 July 2024
The deadline to apply for this project is 5pm on Friday 5th April 2024. To make an application, please go to the following link and complete the application form: https://forms.office.com/e/pMgea0dAHv. To find this project in the application form, please filter ‘school of project applying to’ and select School of Humanities