We have all heard the drumbeat that “AI” systems like LLMs will have enormous implications for researchers and ethnographers. Products like ChatGPT promise a universal interface that may augment or replace human thought, and a growing number of LLM-driven tools promise faster and better insights. Meanwhile, it is challenging to assess the actual capabilities of these tools with respect to the range of data and analytical techniques we use to create multidimensional insights about social worlds.
In this talk, Peter Leonard, an expert in computational humanities who advises faculty across Stanford University on computational methods and data practices, orients us to the nature of these tools and their the growing value and limits for research. He explores the kind of information and collections previously isolated from data-mining practices that are now becoming computationally tractable, and new capabilities such as transcribing human speech, summarizing and analysing themes in transcriptions and other texts, identifying un-described images, and multi-modal analysis of complex cultural artefacts. Peter also discusses concerns around mainstream commercial models that researchers should be aware of; alternative, open-source tools; and useful applications of large, small, and custom models. Following Peter’s presentation, discussants explore specific examples and use cases.
Speaker
Peter Leonard is the Assistant University Librarian for Research Data Services at Stanford University, where he leads teams supporting GIS, data acquisition & analysis, data curation, and AI inference and modeling. He previously served as the Director of the Digital Humanities Lab at Yale University Library and taught in the Department of Statistics and Data Science at Yale. He holds a PhD in Scandinavian Literature from the University of Washington in Seattle, served as a Fulbright Fellow at Uppsala University, and was an Aspen Ideas Festival Scholar.
Discussants
Silvana di Gregorio, PhD is Product Research Director and Head of Qualitative Research at Lumivero, the developers of NVivo, a software program for qualitative and mixed methods research. She is a sociologist and former academic with a PhD in Social Policy from the London School of Economics. She has been training, consulting, and publishing about qualitative data analysis software since 1995. For 16 years she had her own training and consulting business, SdG Associates. She is author of Voice to Text: Automating Transcription and Using Web 2.0 tools for Qualitative Analysis, and co-author with Judith Davidson, Qualitative Research Design for Software Users and Qualitative Research and Technology: In the Midst of a Revolution, and co-author with Linda Gilbert and Kristi Jackson, Tools for Qualitative Analysis. She is part of the Product Team at Lumivero.
Tom Hoy is a Partner at Stripe Partners. He has spent 15 years advising some of the world’s leading organisations on strategy and innovation. Tom’s expertise lies in applying social science theory to unlock concrete business and product challenges. The frameworks and concepts developed by Tom’s teams guide the activity of clients including Apple, Google and Spotify. His work has been covered by the Financial Times and The Guardian. Prior to co-founding Stripe Partners, Tom was a leader in the social innovation field, growing a hackathon network in South London to several hundred members to address local causes. Tom holds a Masters in International Relations from the LSE.
