Generative AI risks and resilience: How users adapt to hallucination and privacy challenges
Purpose: This study examines two central risks affecting continued use of generative AI (GenAI)—AI hallucinations and privacy concerns—and explores how protective behaviors serve as adaptive mechanisms to mitigate these risks. Design/methodology/approach: Drawing on Protection Motivation Theory, the...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-09-01
|
| Series: | Telematics and Informatics Reports |
| Subjects: | |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2772503025000362 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Purpose: This study examines two central risks affecting continued use of generative AI (GenAI)—AI hallucinations and privacy concerns—and explores how protective behaviors serve as adaptive mechanisms to mitigate these risks. Design/methodology/approach: Drawing on Protection Motivation Theory, the study tests a risk-adaptive GenAI use model using survey data from 789 users recruited via a Prolific panel. Structural equation modeling is employed to analyze direct and moderating effects. Findings: Privacy concerns negatively influence user attitudes while positively predicting both personal and system-level protective behaviors. Hallucination risk is similarly associated with negative attitudes but positively predicts information verification. Notably, only information verification significantly moderates the link between attitude and continuance intention. Originality: The study extends Protection Motivation Theory to the GenAI context by developing and validating new constructs—hallucination risk and system-level privacy-protective behaviors. It reveals how different types of risk trigger distinct behavioral adaptations, highlighting the dual role of risk as both a deterrent and catalyst in GenAI adoption. Practical implications: The study offers insights that risk perceptions may not deter continued use of GenAI when users perform risk protective behaviors. By highlighting which protective behaviors enhance continued use, the findings inform risk-mitigation strategies for developers, educators, and regulators. |
|---|---|
| ISSN: | 2772-5030 |