With the advent of generative AI, new possibilities have emerged for the social sciences and humanities. This is felt nowhere more clearly than in the traditionally more qualitatively oriented disciplines where large language models are annotating interview data, extracting information from archived records, writing up analyses based on field notes, providing input for participatory design processes, or being fine tuned / prompt ingested to act like synthetic informants, conduct interviews, or model real world interactions.

Fields like history, anthropology, or comparative literature, which are all firmly rooted in qualitative traditions, now have yet another impetus to rethink their relationships with data science and the past year has seen a wealth of interesting experiments coming out of these and related disciplines. What are the challenges and opportunities arising from these new computational yet deeply qualitative research practices?

Program

8 speed talks (30 mins):

Leveraging Gen AI in Translating Immigrant Oral Histories
Roman Jurowetski (Ass. Prof., AAUBS)

Exploring LLM self-consistency
Mathieu Jacomy (Ass. Prof., AAU, TANTlab)

Synthetic interlocutors in ethnographic research
Laura Kocksch

Citizen AI and AI as citizens – an experiment with political chatbots from Folkemødet
Johan Irving Søltoft

Coming of Age in Stable Diffusion
Anders Kristian Munk (Prof., DTU)

LLMs as companions in controversy mapping
Matilde Ficozzi (Res. Ass., AAU, TANTlab)

Measuring the semantic wildness of wicked problems
Sofie Burgos Thorsen (Senioranalytiker, INVI)

Collaborative literature
exploration with full-text embeddings
Lasse Uhrskov Kristensen (PhD stud., DTU)

Workshop (45 mins)
Together, we will explore an experimental exhibition design created by researchers to engage the public in technical conversations about AI. The exhibition leverages LLMs to summarize the scientific literature and create politically slanted chatbots that prompt discussions with the visitors.

Wrap-up (15 mins)