5: Deep Learning and Culture
Louis Ravn, Anna Schjøtt Hansen, Dasha Simons, Paula Helm and Tobias Blanke
The recent successes of deep learning technologies, including generative artificial intelligence systems, are premised upon deeply cultural elements: training datasets reflecting various forms of cultural activity (language, art, etc.), new cultural techniques like prompting and fine-tuning, massive material investments underpinned by cultures of innovation, and technoscientific practices of AI development. Flipping around this equation, it is equally true that contemporary culture, broadly conceived, has become increasingly suffused by the operations of deep learning, whether in the fields of media and arts, work, education, medicine, or even warfare. In short, the contemporary relationship between deep learning and culture is one of co-constitution – or even entanglement: they crucially depend on and shape each other.
Despite these complex entanglements, scholarship has depicted deep learning as unidirectionally homogenizing culture. While we acknowledge this, we also want to foreground the productive potentials emerging at the encounter between deep learning and culture, potentials that warrant more research to put them onto solid empirical and theoretical foundations. We ask: How can the charged encounter between deep learning and culture help us think about both in new and surprising ways? What happens when deep-learning systems process cultural materials replete with sociopolitical frictions? To what extent can cultural materials be represented by the operations of deep learning systems? How can the methods and epistemologies of deep learning advance critical accounts of culture?
To investigate these questions, we invite contributions from scholars researching the nexus of deep learning and culture drawing from the fields of STS, critical AI studies, digital humanities, media studies, anthropology of technology, and related fields. Specifically, we are interested in contributions that present case studies, methodological approaches, and theoretical frameworks that help us problematize the nexus between deep learning and culture in new ways. We explicitly welcome interdisciplinary approaches to this. Thus, we warmly invite the following contributions (but not limited to):
- Case studies examining the frictions that emerge at the intersection of deep learning and culture
- Methodologies to investigate frictions at the nexus of deep learning and culture (qualitative, computational, quali-quantitative, etc.)
- Theoretical approaches foregrounding the entanglements between deep learning and culture (e.g., to what extent is “Deep Culture” a productive framework?)
- Empirical investigations into the ways in which deep learning and culture are co-shaping each other and what this implies for different groups of people
Keywords: deep learning, culture, artificial intelligence, data, frictions, deep culture