ScienceGuardians

ScienceGuardians

Did You Know?

ScienceGuardians protects the academic community

Task–Technology Fit Leads to Conflict: The Double-Edged-Sword Effect of Generative Artificial Intelligence on Scientific Creative Performance in Humanities and Social Sciences Research

Authors: Pengbo Qian,Xingzheng Xie
Journal: International Journal of Human–Computer Interaction
Publisher: Informa UK Limited
Publish date: 2025-7-14
ISSN: 1044-7318 DOI: 10.1080/10447318.2025.2526580
View on Publisher's Website
Up
0
Down
::

The abstract states that TTF “generates a paradox by enhancing AI literacy while increasing AI dependence” (emphasis added). However, the path coefficients reveal a stark asymmetry: TTF→AID (β=0.579, p<0.001) is nearly twice as strong as TTF→AIL (β=0.325, p<0.001).

Does this not suggest that TTF’s “enhancing” effect on AI literacy is substantially weaker than its “inhibiting” effect through dependence? How can the authors justify framing this as a balanced “paradox” when the data shows a dominant negative pathway? The “Empowerment-Inhibition Paradox” appears empirically lopsided.

All Replies

Viewing 1 replies (of 1 total)

1 week, 5 days ago

In Table 4, the Heterotrait-Monotrait (HTMT) ratio for AI Literacy and AI Dependence is 0.900. This is dangerously close to or exceeds the conservative threshold of 0.85 (or 0.90). When HTMT is this high, it suggests that these two constructs may not be empirically distinct. This poses a fundamental threat to your model, as you are treating them as separate mediators with opposing effects. How can you justify that AI Literacy and AI Dependence are truly distinct constructs given this borderline discriminant validity? If they are essentially the same, your “double-edged sword” mechanism collapses.

There appears to be a fundamental logical contradiction in the mediation analysis results reported in Table 6 and the accompanying text. In H4b, the path is TTF -> AIL -> AUT -> CP. However, the text states the mediating effect of AI literacy on automation is non-significant. This contradicts the logic of your paper, as you specifically hypothesized that AI literacy negatively influences automation. If AI literacy does not significantly reduce automation, how can you still argue it is a critical mechanism for fostering creativity? The non-significant path undermines a core pillar of your model.

Viewing 1 replies (of 1 total)

  • You must be logged in to reply to this topic.