AI Trust: Can Explainable AI Enhance Warranted Trust?

Table 4

Results of ANOVA for the risk and performance interaction on the dependent variables.

Dependent variable	High performance		Low performance			value
	High risk	Low risk	High risk	Low risk
MDMT					10.02	0.002	0.047
Direct subjective trust					5.99	0.015	0.028
Behavioural trust					0.130	0.72	0.001
AI understandability					4.33	0.039	0.021
XAI quality					6.51	0.011	0.031

indicates significant results at the significance level of 0.05. In all self-reported measures, when the performance of the AI system is low, trust levels are higher when the risk is low compared to a high-risk scenario.