Abstract
Since the public launch of ChatGPT (3.5) in November 2022, the world has been
in awe and fear of artificial intelligence. In this paper, ChatGPT, an example of
conversational AI, is investigated from an ethnomethodological and
conversational perspective. How our societies cast and speak of AI has a bearing
on how we perceive and interact with conversational AI. To understand how
ChatGPT is cast in the social world is perspectivised through the agency question
and the notion of anthropomorphism. Further, select articles and online texts are
drawn upon to qualify our understanding of ChatGPT as an example of
conversational AI. To shed light on practices of ChatGPT-interaction, select
human-ChatGPT interactions are analyzed. The analyses show that the AI
displays conversational traits, but signs of demonstrating human-like intelligence
and sentience remain an open question for future studies, despite a consistent
user tendency to anthropomorphize the AI. Conversation Analysis (CA)
scholarship urgently calls for the use of CA in the design of future conversational
AI. This author agrees and proposes distinct development paths for further
research within the realms of ethnomethodology and CA to qualify our design,
understanding of, and interaction with conversational AI.
in awe and fear of artificial intelligence. In this paper, ChatGPT, an example of
conversational AI, is investigated from an ethnomethodological and
conversational perspective. How our societies cast and speak of AI has a bearing
on how we perceive and interact with conversational AI. To understand how
ChatGPT is cast in the social world is perspectivised through the agency question
and the notion of anthropomorphism. Further, select articles and online texts are
drawn upon to qualify our understanding of ChatGPT as an example of
conversational AI. To shed light on practices of ChatGPT-interaction, select
human-ChatGPT interactions are analyzed. The analyses show that the AI
displays conversational traits, but signs of demonstrating human-like intelligence
and sentience remain an open question for future studies, despite a consistent
user tendency to anthropomorphize the AI. Conversation Analysis (CA)
scholarship urgently calls for the use of CA in the design of future conversational
AI. This author agrees and proposes distinct development paths for further
research within the realms of ethnomethodology and CA to qualify our design,
understanding of, and interaction with conversational AI.
| Original language | English |
|---|---|
| Title of host publication | Proceedings of the 2023 Aarhus International Conference on Voice Studies (SEFOS) |
| Editors | M. Hejná, J. Keldgaard-Christiansen, M. Eaton, M. Clasen, Z. Boyd, O. Niebuhr |
| Number of pages | 8 |
| Publisher | Sciendo (DeGruyter) |
| Publication date | Jun 2024 |
| Pages | 84-91 |
| DOIs | |
| Publication status | Published - Jun 2024 |
Keywords
- Artificial intelligence, chatGPT, agency, anthropomorphism, digital interaction, communication design