For students, privacy (Answer-AI) and explainability (Emotion-AI) each positively influenced intention to use indirectly through perceived global utility¶

Claim · All claims
Evidence · 1 study (2 entries) · causal r2 · q3
Currency · newest study 2025 · not yet checked against its sources · not reviewed by a person

Subclaims¶

q3 i? In the student Answer-AI model, privacy had a significant positive impact on global utility (β = 0.2987, p = 0.0439) with a significant indirect effect on intention to use through GU (Effect = 0.0944, BootLLCI = 0.0022, BootULCI = 0.2283). → Alexander John Karran 2025 q3 i? In the student Emotion-AI model, explainability had a significant positive impact on global utility (β = 0.2787, p = 0.0259) with a significant indirect effect on intention to use through GU (Effect = 0.0861, BootLLCI = 0.0039, BootULCI = 0.2018). → Alexander John Karran 2025 (2)

Evidence¶

Alexander John Karran 2025¶

Alexander John Karran, Patrick Charland, Joé Trempe-Martineau, Ana Ortiz de Guinea Lopez de Arana, Anne-Marie Lesage, Sylvain Sénécal & Pierre-Majorique Léger. (2025). Multi-stakeholder perspective on responsible artificial intelligence and acceptability in education. npj Science of Learning 10:44. https://doi.org/10.1038/s41539-025-00333-2

q3 · i? · causal · r2

Mediation analysis of student responses to the Answer-AI chatbot vignette. The direct effect of privacy on intention to use was not statistically significant (β = 0.0369, p = 0.7346); the indirect effect through GU was statistically significant. GU significantly impacted ITU (β = 0.3159, p = 0.0007).

"privacy had a significant positive impact on GU (β = 0.2987, p = 0.0439), indicating that higher levels of privacy were associated with higher levels of both personal and overall benefit"

Alexander John Karran 2025 (2)¶

Alexander John Karran, Patrick Charland, Joé Trempe-Martineau, Ana Ortiz de Guinea Lopez de Arana, Anne-Marie Lesage, Sylvain Sénécal & Pierre-Majorique Léger. (2025). Multi-stakeholder perspective on responsible artificial intelligence and acceptability in education. npj Science of Learning 10:44. https://doi.org/10.1038/s41539-025-00333-2

q3 · i? · causal · r2

Mediation analysis of student responses to the Emotion-AI emotional-support vignette. The direct effect of explainability on ITU was not statistically significant (β = -0.0429, p = 0.7187); the indirect effect through GU was statistically significant. GU significantly impacted ITU (β = 0.3089, p = 0.0039).

"explainability had a significant positive impact on GU (β = 0.2787, p = 0.0259), indicating that higher levels of explainability are associated with higher levels of personal and overall benefit"

Discussion¶

Related Claims¶

• For students, AI agency reduced intention to use Correct-AI indirectly through lower perceived global utility and justice, with no direct effect — related
• For teachers, explainability (Correct-AI) and both agency and explainability (Answer-AI) influenced intention to use indirectly through global utility and perceived usefulness — related
• For teachers, higher privacy levels lowered confidence regarding an AI supporting student emotional health, indirectly reducing intention to use (Emotion-AI) — related
• Higher AI agency is associated with lower intention to use AI in education, while transparency, explainability, and privacy do not directly predict intention to use — related