Cross-cultural adaptation and psychometric evaluation of the artificial intelligence learning intention scale (AILIS) among medical sciences students: a methodological study

Clinical Snapshot

Key Findings

• P Value: Not explicitly reported; CFA chi-square ratio χ²/df = 3.42 (acceptable threshold <5.0)

• Effect Size: Factor loadings >0.40 across all items; total variance explained = 44.55% (EFA)

• Primary Outcome: Four-factor structure confirmed (Epistemic Capacity, Psychological Attitudes, Facilitating Environments, Psychological and Behavioral Outcomes) with acceptable construct validity and reliability in the Persian AILIS

• Nnt Or Sensitivity: Psychometric indices: RMSEA = 0.078 (acceptable <0.08), CFI = 0.932 (acceptable >0.90); Cronbach's alpha 0.854–0.904; McDonald's omega 0.854–0.899; ICC 0.750–0.860 (good to excellent test-retest reliability)

• Confidence Interval: Not reported in the abstract for any psychometric index

Clinical Application

High feasibility for educational research and curriculum evaluation purposes. The instrument is a self-administered survey requiring no specialised equipment or clinical infrastructure. Administration burden is low. Utility is primarily for medical education researchers, curriculum developers, and health professions educators rather than direct clinical practitioners. The Persian AILIS is not directly applicable to Australian clinical or educational settings without English-language adaptation and re-validation. However, the study is highly relevant to Australian medical education policy given the increasing integration of AI into medical curricula (consistent with AMC accreditation standards and RACGP curriculum frameworks). Australian medical educators and researchers could use this validated methodology as a template for developing or adapting equivalent instruments for Australian medical students. The TGA's growing regulatory engagement with AI-enabled medical devices and the Australian Digital Health Agency's AI strategy further underscore the need for validated tools to assess AI learning readiness in Australian health professional education. No PBS or TGA regulatory implications apply directly to this measurement instrument. Medical sciences students in Persian-speaking educational contexts; the validated instrument is suitable for use in Iranian undergraduate and postgraduate medical, nursing, and allied health programmes. The methodology is transferable to other cultural contexts requiring local adaptation.

Abstract

BACKGROUND: This study aimed to adapt and psychometrically evaluate the Artificial Intelligence Learning Intention Scale (AILIS) among Iranian medical sciences students. METHODS: This methodological psychometric study was conducted among 800 medical sciences students at Babol University of Medical Sciences, Iran. The AILIS was translated and culturally adapted into Persian using a standardized cross-cultural adaptation process. Construct validity was examined through exploratory factor analysis (EFA) and confirmatory factor analysis (CFA). Reliability was assessed using Cronbach's alpha, McDonald's omega, and the intraclass correlation coefficient (ICC). Convergent and discriminant validity were evaluated using composite reliability (CR) and average variance extracted (AVE). Measurement invariance across gender and structural stability were assessed using multi-group confirmatory factor analysis and exploratory graph analysis (EGA). RESULTS: EFA supported a four-factor structure consisting of Epistemic Capacity, Psychological Attitudes, Facilitating Environments, and Psychological and Behavioral Outcomes, explaining 44.55% of the total variance, with factor loadings greater than 0.40. Confirmatory factor analysis indicated an acceptable model fit, supporting the proposed factor structure (χ²/df = 3.42, RMSEA = 0.078, CFI = 0.932). Internal consistency and test-retest reliability were satisfactory, with Cronbach's alpha ranging from 0.854 to 0.904, McDonald's omega from 0.854 to 0.899, and ICC from 0.750 to 0.860. Composite reliability values were adequate, while AVE values indicated acceptable but partially limited evidence of convergent validity. Discriminant validity was generally supported. In addition, multi-group analysis demonstrated measurement invariance across gender, and exploratory graph analysis confirmed the structural stability of the four-factor model. CONCLUSION: The Persian AILIS is a valid and reliable instrument for assessing AI learning intention among medical sciences students. Beyond its psychometric robustness, the instrument may support curriculum planning, evaluation of AI-related educational interventions, and evidence-based decision-making in medical sciences education by providing a standardized measure of students' intentions to engage with AI learning.

References

1. 1.Taghiparast, A. H., Mahmoudian, A., Ghaemi-Amiri, M., & Ghaffari, F. (2026). Cross-cultural adaptation and psychometric evaluation of the artificial intelligence learning intention scale (AILIS) among medical sciences students: a methodological study. Medical Education Online. https://doi.org/10.1080/10872981.2026.2690326