AI & Society
Expected usefulness, ease of use, and habit are linked to more chatbot use for knowledge sharing among students, while perceived threats are linked to less use, a 2023 survey reports
In a 2023 study published in IEEE Transactions on Engineering Management, researchers surveyed 447 students about using AI-based chatbots for knowledge sharing. Students who expected a chatbot to be useful, found it easy to use, and had a habit of using it were more likely to report using it to share knowledge. Students who felt some perceived threat from the chatbots were less likely to report that use.
What were the main factors?
The study used a symmetric statistical method called partial least squares structural equation modeling. In that analysis, performance expectancy, effort expectancy, and habit showed positive links to chatbot use. Performance expectancy means expecting the chatbot to be useful for sharing knowledge. Effort expectancy means expecting it to be easy to use. Perceived threats showed a negative link.
The same analysis did not show an impact for social influence, facilitating conditions, or hedonic motivation. Social influence is what other people around a student do or approve of. Facilitating conditions are available support and resources. Hedonic motivation is enjoyment.
What did the second analysis add?
The researchers also used fuzzy set qualitative comparative analysis, an asymmetric approach that examines combinations of factors rather than one factor at a time. In that second analysis, all the factors might play a role in shaping chatbot use, including the three factors the standard analysis did not support.
FAQ
Did the study prove that these factors cause chatbot use?
No. The study used a questionnaire survey of 447 students. It reports links between attitudes and reported chatbot use, not a controlled demonstration of cause.
What does the second method add?
The second method looks at combinations of factors. It suggested that factors the first method did not link to chatbot use, such as social influence and enjoyment, might still matter for some students.
What practical use do the authors see?
The authors say the findings offer implications for universities, instructors, and chatbot developers who want to make better decisions and promote chatbot use.
Limits and what's next
This is one questionnaire survey of 447 students. The abstract does not state where the students were located or how they were selected, so the findings describe this group only. The study used students' survey answers, which reflect what they said about chatbot use rather than a direct record of activity. The abstract reports no independent replication of the results.
Source: Mostafa Al‐Emran et al., “Determinants of Using AI-Based Chatbots for Knowledge Sharing: Evidence From PLS-SEM and Fuzzy Sets (fsQCA),” IEEE Transactions on Engineering Management, 2023.