Hedonic vs Utilitarian No.4

1. Title (Reference):

   Chattaraman, V., Kwon, W. S., Gilbert, J. E., & Ross, K. (2019). Should AI-Based, conversational digital assistants employ social-or task-oriented interaction style? A task-competency and reciprocity perspective for older adults. Computers in Human Behavior, 90, 315-330.

2. Objective:

   • investigate the answer to such questions in the context of social- and task-oriented conversation styles of virtual shopping assistants for the growing online community of older adults.
   • 2 (digital assistant interaction style: (social-vs. task-oriented)× 2 (user Internet competency: low vs. high)×2 (user exchange modality: text vs. voice) between-subjects design.

3. Agent Type: virtual shopping assistants/web-based male embodied agent,

4. Theme of conversation (domain): informal/formal conversation, in context to an online shopping task

5. Result / Implication

   • the task-oriented reciprocity (users focusing on the task when the computer is task-oriented) is augmented for low-competency users, while the social reciprocity (users treating a socially-oriented computer system in a social way) is intensified for high-competency users.

   • it is noteworthy that these competency-based reciprocity effects emerged significantly only in the user outcome dimensions that matched the domain of the reciprocity (i.e., significant social reciprocity effects only for social outcomes and significant task-oriented reciprocity effects only for task-related [functional and cognitive] outcomes).

   • context-aware digital assistant conversation design based on the contingency between the interaction/task goals (e.g., social/relational vs. functional/instrumental goals) and the user’s task-competency

6. Memo:

   • the purpose of the current study was to examine whether social- versus task-oriented conversation styles of a virtual shopping assistant differentially benefits low versus high Internet competency older adults with respect to a) social outcomes (perceived interactivity and trust), b) cognitive outcomes (perceived information load), c) functional outcomes (perceived self-efficacy, perceived ease of use, and perceived usefulness), and d) behavioral intent outcomes (website patronage intent) in context to online shopping
   • Cerekovic et al. (2017) argued that these reciprocal behaviors are more evident when the user was more extroverted(big five) and agreeable (big five) by nature.
   • the user’s self-disclosure was greater when the computer first self-disclosed information about itself (vs. not).
   • user’s work ethic was reciprocal to the computer’s helpfulness
   • all of the above findings on the reciprocity principle essentially suggest that humans socially respond to computers by trying to ‘reciprocate’ or ‘match’ the computer’s interaction or behavior
   • users interacting with a task-oriented digital assistant may reciprocate by being more task-oriented; whereas those interacting with a social-oriented digital assistant may reciprocate by being more social-oriented. *Task- oriented conversation is focused on assisting users in achieving functional goals (e.g., completing the online shopping task), which may ‘match’ the needs of low task-competency users; whereas social-oriented dialog, which aims at socioemotional and relational goals, may better align with the mental orientation of high task-competency users whose functional goals can be easily met without assistance
   • extensive and extended inclusion of social comments within a task-oriented pedagogical context does not benefit user interactivity (Veletsianos, 2012). social conversation in task-oriented contexts can produce mixed results depending the amount and nature of social comments. Limited inclusion of social dialog seems to be the best approach
   • social dialog may be more beneficial for building trust than task-oriented dialog, irrespective of user characteristics such as task-competency
   • the effectiveness of social-versus task-oriented conversation of digital assistants emerges when their interaction style matches the user characteristics (i.e., social-oriented interaction for an extroverted user) or the context (i.e., social-oriented interaction for shopping for credence goods).
   • small talk functioned as a prerequisite in establishing trust for extroverted users
   • conversational digital assistants should model 1) the desired user outcome/goal or the need state of the user and 2) the task competency of the user
   • (Bickmore and Cassell 2005) small talk in an embodied real estate conversational agent led to greater trust as compared to task dialog for extrovert personality users