This study examines humorous interactions with intelligent personal assistants (IPA/IPAs). The respective IPAs include Google Assistant, Amazon Alexa, Microsoft Cortana, and Apple Siri. Many discussions of a wide-ranging review helped to define the goal to classify user utterances, IPA responses, and user ratings of IPA responses.
Resulting user ratings and joke descriptions from online diaries and paper questionnaires. The content analysis method led the examination and categorization of this data.Findings suggest that users are disposed to question and test the IPAs. These tests attempt to help the user determine inherent characteristics and perspectives. In essence, users are interested in the personalities of these systems. Joke requests also appear with a high frequency. These requests are met with pre-programmed replies that users generally find funny.The published datasets not only validate but also expand upon the initial classification of humorous utterances. The findings can be applied in an effort to immediately improve IPA performance to alleviate user experiences. The study also supports the long-term development of IPA personas and algorithms.
Research team: Dr. Irene Lopotovska, Katherine Curran, Armando Garcia, Mary Mann, Shannon Mish, Alexandra Srp, Sydney Stewart, Alanood Al Thani, Wanyi Wang
Special thanks to Alice Griffin
#infoshow Presenters: Katie Curran, Armando Garcia, Shannon Mish
Supported by: Wanyi Wang