Abstract
This talk is based on a recent article that introduces the concept of stochastic remembering to better understand the distributed mnemonic agency involved the use of generative AI tools like ChatGPT. The main outcome of a TAIGA micro project called: Artificial Intelligence and Social Memory - Critical Explorations of AI’s Implications for Societal Remembrance that took place 2023, the article brings critical AI studies into greater dialogue with the interdisciplinary field of memory studies.

In the article we use two prompt engineering techniques to critically examine the mnemonic agency of ChatGPT via the texts that it generates. Specifically, these techniques – step-by-step prompting and chain of thought reasoning – are experimentally applied to understand how ChatGPT shapes how we remember historical activists. This experiment highlights how hegemonic forms of memory influence the data on which AI chatbots like ChatGPT are trained and underlines how stochastic patterns affect how humans and AI chatbots collectively remember the past. Humans and AI chatbots prompt each other to remember. In conclusion, we argue that AI chatbots are a new kind of mnemonic actor that, in interaction with users, stochastically renders a probabilistic past.

The article: https://brill.com/view/journals/mesr/1/2/article-p209_002.xml

The microproject: Artificial Intelligence and Social Memory - Critical Explorations of AI’s Implications for Societal Remembrance

About the speakers

Samuel Merrill is Associate Professor at Umeå University’s Department of Sociology and Centre for Digital Social Research (DIGSUM). He specializes in digital and cultural sociology and his research interests concern, among other things, the intersections between memory and digital technology, social media platforms, and AI systems.

Rik Smit is an Assistant Professor at the Centre for Media and Journalism Studies at the University of Groningen, the Netherlands. He teaches and does research within the field of memory studies, digital media, algorithmic culture and critical ai studies. His research has appeared in a range of journals and books, including New Media & Society, Convergence and Memory Studies.

Thomas Smits is Assistant Professor of Digital History & AI at the University of Amsterdam. His work is centered on modern visual (news) culture and is located at the intersections of digital humanities and social sciences. He both critiques ai and applies it in his research.