By: Haley Anne Schwartz
After spending another summer melting in Barcelona, I traded in my office (where I don't know how to work the aircon) at the Universitat de Barcelona, for a flat in Edinburgh (where I couldn't figure out how to work the heat) for a research stay at the University of Edinburgh. While this wasn't my first rodeo with the University of Edinburgh (I previously completed my MSc here in 2020), it was my first rodeo working with Generative AI tools. This stay allowed the chance to expand on my current doctoral research in quantitative textual analysis applied to archaeological heritage discourse into the great big world of Generative AI.
Specifically, cultivating a project with the objective of analysing archaeological heritage texts produced by ChatGPT-4. Under the supervision of Dr. Chiara Bonacchi, the project has been formally shaped into assessing the impact of prompt engineering on the quality of both 1] responses from ChatGPT-4, categorized as newly produced knowledge and 2] said knowledge in the context of user consumption. This project uses a mixed-methods approach to analyse responses from interviewees and ChatGPT-4. Case studies are focused on neolithic monuments within Scotland and England, prompted in a heritage tourism context. As this research is still very much in the 'early days' phase, there is not much to report on beyond some insights into both ChatGPT-4 and potential directions of archaeological heritage discourse research.
During the stay, I had additional opportunities to audit MSc intensive courses taught by Dr. Bonacchi and Dr. Arturo Rey da Silva, attend CDCS trainings and workshops, and participate in some "In Theory" sessions run by PhD's in the HCA. While the stay is over, I am grateful to have left Edinburgh having met and learned from so many interesting people, progressing with this research, and filling my suitcase with an inordinate amount of Christmas puddings and millionaire shortbread.