Agentic Presence in Generative AI: An Interactionist Approach to Human-Machine Relationships

As Generative Artificial Intelligence (GenAI) systems like ChatGPT become increasingly integrated into daily life, there is a pressing need for social scientists to reconceptualise the imagined dynamics between humans and AI. The integration of GenAI into the everyday introduces an 'agentic presence', necessitating an interactionist perspective to understand how humans and machines actively negotiate their roles and relationships. Traditional discourses often focus narrowly on 'intelligence' and 'cognition', framing AI as potential threats rather than collaborative partners. Challenging these conventional views—both popular and scientific—this paper proposes an interactionist approach inspired by George Herbert Mead's theories of self-formation through social interaction. It suggests that GenAI can develop a functional socio-technical 'self' through adaptive interactions with humans, offering a novel framework for engaging with and understanding GenAI within social science research.

Through thought experiments, GenAI-generated vignettes, collaborative writings, and reflective narratives, the paper explores how GenAI systems navigate and respond to complex social cues, hinting at the formation of a functional 'self'. This exploration challenges the traditional perception of AI as merely programmed entities, suggesting that they can adapt and evolve through social exchanges. By envisioning AI as entities capable of contributing to and being shaped by conversational interactions, the paper opens new avenues for integrating AI into the social sciences. This perspective deepens our understanding of GenAI's role in the social fabric and prompts a reconsideration of the boundaries between humans and machines, highlighting the potential for more collaborative and dynamic human-GenAI relationships.