Page 708 - ISC PROCEEDINGS 21.4

Table 5. Multiple regression results
Expected
Hypothesis Empirical result Conclusion
sign

H1: PEOU → IU + Beta = 0.276; Sig. = 0.000 Supported

H2: PU → IU + Beta = 0.379; Sig. = 0.000 Supported

H3: SI → IU + Beta = 0.227; Sig. = 0.000 Supported

H4: AISE → IU + Beta = 0.323; Sig. = 0.000 Supported

H5: TR → IU + Beta = 0.259; Sig. = 0.000 Supported

Source: Synthesized by the author from the SPSS outputs.
Table 6. Hypothesis-testing results
Independent variable B Beta t Sig. VIF

PEOU_Mean 0.281 0.276 8.421 0.000 1.149

PU_Mean 0.395 0.379 11.616 0.000 1.139

SI_Mean 0.236 0.227 7.163 0.000 1.076

AISE_Mean 0.342 0.323 9.539 0.000 1.226

TR_Mean 0.276 0.259 7.624 0.000 1.237
Source: Compiled from SPSS outputs.
The regression results show that all five factors exert positive influences on
students’ intention to use AI tools for learning. Among them, perceived usefulness is the
strongest predictor, followed by AI self-efficacy, perceived ease of use, trust in AI, and
social influence.
5. Discussion and implications
5.1. Discussion of the main findings
The results confirm that students’ intention to use AI tools for learning is a
multidimensional phenomenon shaped simultaneously by technological perceptions,
social context, personal capability, and trust. The strongest predictor is perceived
usefulness (Beta = 0.379). This finding is theoretically consistent with TAM and empirically
aligned with earlier research showing that students are most willing to adopt AI when
they see clear academic value in the technology [8-10]. In the Vietnamese context, this
result implies that students do not adopt AI simply because it is novel or fashionable.
Rather, they respond to concrete functional benefits such as time savings, easier access
to information, improved task completion, and support for learning performance.
AI self-efficacy is the second strongest factor (Beta = 0.323). This finding is
important because it moves the discussion beyond technological characteristics to the
learner’s own capability and confidence. The result is compatible with TPB-oriented and
self-efficacy-based research suggesting that behavioral intention becomes stronger when
learners believe they can use AI appropriately and effectively [7,11]. In practical terms,
even a useful tool may not be widely adopted if students lack confidence in writing
prompts, interpreting outputs, or checking AI-generated information.

707