Page 649 - ISC PROCEEDINGS 21.4

Table 1. Measurement model results
Construct Items Loadings Cronbach’s α CR AVE
Perceived Ease of Use (PEOU) 4 0.74–0.88 0.88 0.91 0.71
Perceived Usefulness (PU) 4 0.76–0.90 0.90 0.93 0.76
Attitude Toward AI (ATT) 3 0.79–0.91 0.87 0.92 0.79
Social Influence (SI) 3 0.73–0.86 0.82 0.89 0.72
Trust in AI (TR) 3 0.77–0.88 0.85 0.90 0.74
AI Literacy (AIL) 4 0.72–0.87 0.86 0.90 0.69
Behavioral Intention (BI) 3 0.82–0.92 0.89 0.93 0.82
Actual Use (AU) 3 0.75–0.89 0.84 0.89 0.73
All constructs demonstrated strong internal consistency and convergent validity.
4.4. Discriminant validity (HTMT)
The Heterotrait–Monotrait (HTMT) ratio was used to assess discriminant validity. All
HTMT values were below the threshold of 0.85, confirming discriminant validity.
Table 2. HTMT matrix
Construct PEOU PU ATT SI TR AIL BI AU
PEOU -
PU 0.71 -
ATT 0.64 0.78 -
SI 0.49 0.55 0.60 -
TR 0.52 0.63 0.66 0.58 -
AIL 0.59 0.68 0.61 0.47 0.54 -
BI 0.60 0.74 0.81 0.70 0.73 0.65 -
AU 0.48 0.59 0.66 0.62 0.64 0.53 0.78 -
All HTMT values < 0.85, indicating satisfactory discriminant validity.
4.5. Structural model assessment
Bootstrapping (5,000 resamples) was conducted to test the significance of path
coefficients.
Model Fit and Predictive Power R²(Perceived Usefulness) = 0.54 R²(Attitude) = 0.62
R²(Behavioral Intention) = 0.71 R²(Actual Use) = 0.61. The model explains 71% variance in
behavioral intention, indicating strong explanatory power.
Path Coefficients
Table 3. Structural model results
Hypothesis Path β t-value p-value Result

H1 PEOU → PU 0.48 9.87 <0.001 Supported
H2 PU → ATT 0.42 8.54 <0.001 Supported
H3 PEOU → ATT 0.21 4.12 <0.001 Supported
H4 ATT → BI 0.35 7.66 <0.001 Supported
H5 SI → BI 0.22 5.01 <0.001 Supported
H6 TR → BI 0.27 6.14 <0.001 Supported
H7 AIL → PU 0.39 8.21 <0.001 Supported
H8 BI → AU 0.78 18.45 <0.001 Supported
All hypothesized relationships were statistically significant.

648