AI attitude surveys need to separate usefulness, self-expression, and concern

Updated Sep 07, 2026

student voicefeedback

An AI attitude score can hide different reasons for a response. Angyang Li and Shuo Wang developed the Functional Attitudes towards Artificial Intelligence scale with Chinese university students. This review uses the publisher's abstract and notes, plus selected sections of an author-uploaded preview; the full paper was not read in its entirety.

What was tested

The exploratory study involved 366 students, followed by confirmatory analysis with 623. The resulting 12-item scale distinguishes utility-knowledge, value-expression and ego-defense: broadly, usefulness, expression of identity and protective concern.

The instrument addresses AI broadly, including applications outside academic work. It should not be presented as a scale specifically validated for generative AI in UK teaching.

A publisher note reports a follow-up involving 266 students. Utility-knowledge and value-expression correlated positively with reported usage frequency; the ego-defense correlation was not statistically significant. These are associations, not evidence that changing attitudes causes adoption. A non-significant result does not establish that concern never relates to use.

What survey teams could explore

Our suggestion is to distinguish the construct being measured before changing a questionnaire. Usefulness, confidence, trust and concern are not interchangeable labels. Copying a few items or adapting them to a new context does not preserve the original instrument's validation automatically.

Pilot wording with local students, check what they understand and decide whether comparisons across groups are justified. Open-text prompts can invite explanations outside predefined response options. Comments may suggest questions for follow-up, but they cannot by themselves identify a causal driver or determine the right intervention.

The student feedback analysis glossary can support a shared local vocabulary. This is an editorial application, not a tested outcome of the study.

FAQ

Does frequent AI use demonstrate trust?

Usage and attitude measures answer different questions. This study does not justify treating frequency as a trust score.

Can UK institutions use the scale without further checks?

Local suitability needs testing, especially if wording, population or the meaning of AI changes.

Reference

Li, A., & Wang, S. (2025, online). Functional attitudes towards artificial intelligence among university students: Development of a scale and influencing factors. Higher Education. Paper source.

Correction, 7 September 2026: Changed prediction to correlation, clarified the broad AI scope, removed an unverified quotation and qualified claims about survey benefits.