What Would Be the Principles for Successful Trollbot Design?

Heidi Vepsäläinen*, Antti Salovaara, Henna Paakki

*Corresponding author for this work

Research output: Contribution to conferencePaperScientific

Abstract

As far as we know, trollbots that would be indistinguishable from humans and would succeed in luring people into endless frustrating conflicts without being recognized as bots do not yet exist in social media. It is though very likely, that there is a desire to design one for malicious purposes. Here we speculate on the idea of designing a successful trollbot for research purposes by using concepts that derive from Conversation Analysis and Natural Language Framework. Based on our ongoing reseach on trolling, we argue that a successful trollbot would need to prevent its interlocutor from reaching their goal in a given context, but at the same time manage to keep the other party expecting that they would be able to reach a common ground at some point.
Original languageEnglish
Publication statusPublished - 2021
MoE publication typeNot Eligible
EventLet's Talk About CUIs: Putting Conversational User Interface Design Into Practice - Virtual, Online
Duration: 8 May 20218 May 2021
http://www.speechinteraction.org/CHI2021/index.html

Workshop

WorkshopLet's Talk About CUIs: Putting Conversational User Interface Design Into Practice
CityVirtual, Online
Period08/05/202108/05/2021
Internet address

Keywords

  • chatbot
  • Conversation Analysis
  • Natural Language Framework
  • trollbot

Cite this