Playfully Extending Cognition: A Thematic Analysis on Collective Memory, AI and Agency

Abstract

As artificial intelligence (AI) increasingly becomes part of everyday life, it mediates collective memory, reshaping how history is narrated and experienced. Large Language Models (LLMs) like ChatGPT act as memory agents, co-constructing collective memory in a social framework, as they have become tools for accessing, and framing the way we experience information [1]. This new construction of collective memory, where information is filtered and presented through the lens of AI’s learned biases, raises questions about historical truth and representativeness [1], [2]. This challenges us to consider how knowledge is perceived through this new, algorithmically shaped, form of accessing historical information, and how users critically engage with these representations mediated by LLMs. Consequently, inviting further questions of how user's agency is enacted and influenced when confronted with this statistically framed form of history and knowledge. 

This thesis forms part of a larger research project investigating connections between AI and collective memory within a playful setting. Within this framework, my thesis focuses on the user’s experience of agency: How do humans challenge, or accept AI-generated historical narratives during playful interactions? How do machine biases, and framings impact users' sense of agency? What does this imply for this new construction of collective memory?

A qualitative research design is implemented, built around the Ludic Method [3] combining play and critical inquiry. 15 participants will first take part in a role-play game with an AI embodying a historical figure, followed by a semi-structured interview designed to explore their reflections and emotional responses. Both, the gameplay interaction and the interview data will be interpreted using a Thematic Analysis to reveal how users interpret, accept or question AI-generated narratives. Building on the theoretical framework of Extended Cognition and the concept of AI as memory agent [1], this thesis investigates how play, the user's perceived, conceived and lived agency [3] and user's reflections affect interactions with AI and therefore this new co-construction of collective memory.

By exploring how LLMs, especially conversational chatbots, reshape not only what is remembered, but also how humans experience their role in the construction and recollection of memory, this study contributes to the broader discussion about collective memory in digital times, agency, and the political and ethical positions of AI's growing influence on how we perceive and interact with knowledge.

References

[1] J. Schuh, “AI As Artificial Memory: A Global Reconfiguration of Our Collective Memory Practices?,” Mem. Stud. Rev., vol. 1, no. 2, pp. 231–255, Oct. 2024. doi: 10.1163/29498902-202400012.

[2] L. Floridi, “AI as Agency Without Intelligence: on ChatGPT, Large Language Models, and Other Generative Models,” Philos. Technol., vol. 36, no. 1, p. 15, Mar. 2023. doi: 10.1007/s13347-023-00621-y.

[3] M. Jahrmann, “Ludics for a Ludic Society: The Art and Politics of Play,” Ph.D. dissertation, Faculty of Arts, Humanities and Business, Univ. of Plymouth, Plymouth, U.K., 2011.