Location

Hilton Hawaiian Village, Honolulu, Hawaii

Event Website

https://hicss.hawaii.edu/

Start Date

3-1-2024 12:00 AM

End Date

6-1-2024 12:00 AM

Description

Enabled with artificial intelligence (AI), intelligent agents in information systems have developed from passive tools that only help in return to user prompts (i.e., reactive help) to intelligent agents that can help without requiring user requests (i.e., proactive help). Yet, it is unclear how users react to these different types of help and whether the task creates or reinforces the users’ identity (i.e., identity-relevance). Against this backdrop, we drew on self-affirmation and identity theory and conducted a vignette-based online experiment (n = 135). Our results show that proactive (vs. reactive) help decreases users’ willingness to accept help because of users’ higher perceived self-threat (i.e., threat to their self-image). Identity-relevance of the task moderates this effect – high (vs. low) identity-relevance causes a greater increase in self-threat through proactive (vs. reactive) help. Our study contributes to a better understanding of help from intelligent agents and their implications for effective human-AI collaboration.

Share

COinS
 
Jan 3rd, 12:00 AM Jan 6th, 12:00 AM

Proactive and Reactive Help from Intelligent Agents in Identity-Relevant Tasks

Hilton Hawaiian Village, Honolulu, Hawaii

Enabled with artificial intelligence (AI), intelligent agents in information systems have developed from passive tools that only help in return to user prompts (i.e., reactive help) to intelligent agents that can help without requiring user requests (i.e., proactive help). Yet, it is unclear how users react to these different types of help and whether the task creates or reinforces the users’ identity (i.e., identity-relevance). Against this backdrop, we drew on self-affirmation and identity theory and conducted a vignette-based online experiment (n = 135). Our results show that proactive (vs. reactive) help decreases users’ willingness to accept help because of users’ higher perceived self-threat (i.e., threat to their self-image). Identity-relevance of the task moderates this effect – high (vs. low) identity-relevance causes a greater increase in self-threat through proactive (vs. reactive) help. Our study contributes to a better understanding of help from intelligent agents and their implications for effective human-AI collaboration.

https://aisel.aisnet.org/hicss-57/cl/machines_as_teammates/4