Chatting without ChatGPT
Time flies, here is a paper that is 10 years old about studies with a conversational agent in the role of a barkeeper playing out different social scenarios
From 2009 to 2013, I participated in the EU-funded project Cyberemotions. We wanted to understand how emotions spread on the Internet. One of the goals was to see whether we could predict confrontations and perhaps avoid them happening altogether. The paper that was published ten years ago related to how one could potentially prevent negative outbursts by employing chatbots that might be deployed to de-escalate tensions in a virtual space. The article, with Marcin Skowron, Mathias Theunis, and Stefan Rank, reported three experiments with an intelligent chatbot. Today, as everybody talks about ChatGPT and similar systems, there is much talk about the answers intelligent systems can provide. We were interested not simply in a game of answering questions, but to model real interactions with different social functions that deal with affective states.
M. Skowron, M. Theunis, S. Rank and A. Kappas, "Affect and Social Processes in Online Communication--Experiments with an Affective Dialog System," in IEEE Transactions on Affective Computing, vol. 4, no. 3, pp. 267-279, July-Sept. 2013, doi: 10.1109/T-AFFC.2013.16.