The Drawback of Insight: Detailed Explanations Can Reduce Agreement with XAI

Explanations don't universally increase trust: The study challenges the common assumption that providing explanations always improves user agreement with AI recommendations. In fact, detailed explanations sometimes reduced agreement, suggesting a one-size-fits-all approach to XAI design is ineffective.

Personality traits significantly influence how users respond to AI: People with higher neuroticism and lower technological comfort actually showed more agreement with AI suggestions when explanations were absent. This indicates that individual psychological differences, not just the quality of explanations, shape how people interact with AI systems.

XAI design should be personalized, not standardized: Rather than always providing explanations, the researchers recommend that XAI systems should adapt based on user personality profiles. For users with high neuroticism and low tech comfort, withholding detailed explanations may actually improve their collaboration with AI.

More information isn't always better for user confidence: The counterintuitive finding suggests that detailed explanations can overwhelm certain users or trigger skepticism, potentially backfiring on adoption goals. This has important implications for how organizations deploy AI systems to diverse user populations.

Responsible AI requires understanding individual differences: Moving toward truly responsible AI means recognizing that users have different needs and psychological profiles. A responsible approach would assess user characteristics before deciding whether and how to present AI explanations.