Abstract: Anthropomorphic effects of GPT-3, a third-generation natural language modelling system, are leaving a lasting impact on public discourse, communication, and education.  This paper investigates conversational aspects of linguistic AI’s ability to construct believable representations of real people.  Engaging conversations with Plato, Queen Min or Princes Diana are now more possible than ever.  The GPT-3 conversational simulacra seem to possess a form of intentionality (a will to be understood).  Social, philosophical and ethical implications of this new generation of conversational simulacra are discussed within a broad historical tradition of the phenomena of identity transfers: from Eastern mythology to the visions of post-human in contemporary critical discourse.

The paper focuses on the possible impact of AI’s mimetic representations on existing notions of identity, agency, immortality. As such, it hopes to bridge the gap between discourses and bring AI closer to Humanities, and Humanities to AI.

Keywords: GPT-3, avatar, chatbots, posthuman, AI