ABSTRACT
Alan Turing proposed a test for the intelligence of machines in 1950 [1]. Despite great efforts, no computer has passed this test so far. Each year, chat bots compete for the Loebner Prize, the first formal instantiation of a Turing Test. No contender was able to fool the jury yet. Major problems of the chat bots are the lack of common knowledge and the logical consistency of a dialogue.
We explore a new approach to chat bots by focusing on non-logical conversation topics: mysticism. The founding books of the major religions are widely acknowledged examples of mystical topics. We selected the New Testament, the Koran and Rigveda as the knowledge base for our conversational robots.
The robots are able to autonomously talk to each other and to humans about their religious believe. Each robot represents a belief, but we do not reveal their convictions. This ambiguity forces observers to follow the actual conversations instead of quickly applying stereotypes.
Supplemental Material
- A. Turing, "Computing machinery and intelligence," Mind, vol. 59, no. 236, pp. 433--460, 1950.Google ScholarCross Ref
- LEGO, "nxt," http://mindstorms.lego.com, 2009.Google Scholar
- C. Bartneck and J. Hu, "Rapid prototyping for interactive robots," in Intelligent Autonomous Systems 8:{the Intelligent Autonomous Systems Conference, Amsterdam,March 2004}. Ios Pr Inc, 2004, p. 136.Google Scholar
- L. Mayor, B. Jensen, A. Lorotte, and R. Siegwart, "Improving the expressiveness of mobile robots," in Proc. of IEEE Int. Workshop on Robot and Human Interactive Communication ROMAN), 2002.Google ScholarCross Ref
- J. Vandevelde, P Geurts, "Robbedoes en de vallei der bannelingen, 1994.Google Scholar
Index Terms
- Mysterious machines
Recommendations
Colossus: Its Origins and Originators
The British Colossus computer was one of the most important tools in the wartime effort to break German codes. Based on interviews and on recently declassified documents, this article clarifies theroles played by Thomas Flowers, Alan Turing, William ...
'Realness' in chatbots: establishing quantifiable criteria
HCI'13: Proceedings of the 15th international conference on Human-Computer Interaction: interaction modalities and techniques - Volume Part IVThe aim of this research is to generate measurable evaluation criteria acceptable to chatbot users. Results of two studies are summarised. In the first, fourteen participants were asked to do a critical incident analysis of their transcriptions with an ...
Comments