Development and Validation of the Artificial Intelligence Learning Intention Scale (AILIS) for University Students

As artificial intelligence (AI) permeates almost all aspects of our lives, university students need to acquire relevant knowledge, skills, and attitudes to adapt to the challenges it poses. This study reports the development and validation of a scale called the Artificial Intelligence Learning Intention Scale (AILIS). AILIS was designed to measure the different factors that shape university students' behavioral intentions to learn about AI and their AI learning. We recruited 907 Chinese university students who answered the survey. The scale is comprised of 9 factors that are categorized into various dimensions pertaining to epistemic capacity (AI basic knowledge, programming efficacy, designing AI for social good), facilitating environments (actual use of AI systems, subjective norms, access to support and technology), psychological attitudes (resilience, optimism, personal relevance), and focal outcomes (behavioral intention to learn AI, actual learning of AI). Reliability analyses and confirmatory factor analyses indicated that the scale has acceptable reliability and construct validity. Structural equational modeling results demonstrated the critical role played by epistemic capacity, facilitating environments, and psychological attitudes in promoting students' behavioral intentions and actual learning of AI. Overall, the findings revealed that university students express a strong intention to learn about AI, and this behavioral intention is positively associated with actual learning. The study contextualizes the theory of planned behavior for university AI education, provides guidelines on the design of AI curriculum courses, and proposes a possible tool to evaluate university AI curriculum. Purpose: The purpose of this study was to develop and validate a scale called the Artificial Intelligence Learning Intention Scale (AILIS). AILIS was designed to measure the different factors that shape university students' behavioral intentions to learn about AI and their AI learning. Methods: We recruited 907 Chinese university students to answer the AILIS survey. Conclusions: The scale is comprised of nine factors that are categorized into various dimensions pertaining to epistemic capacity (AI basic knowledge, programming efficacy, designing AI for social good), facilitating environments (actual use of AI systems, subjective norms, access to support and technology), psychological attitudes (resilience, optimism, personal relevance), and focal outcomes (behavioral intentions to learn AI, actual learning of AI). Reliability analyses and confirmatory factor analyses indicated that the scale has acceptable reliability and construct validity. Structural equational modeling results demonstrated the critical role played by epistemic capacity, facilitating environments, and psychological attitudes in promoting students' behavioral intentions and actual learning of AI. Implications: Overall, the findings revealed that university students express a strong intention to learn about AI, and this behavioral intention is positively associated with actual learning. The study contextualizes the theory of planned behavior for university AI education, provides guidelines on the design of AI curriculum courses, and proposes a possible tool to evaluate university AI curriculum. Limitations: One key limitation is the adoption of convenience sampling. In addition, this study could be further enriched by the collection of qualitative data. A third limitation is the cross-sectional nature of the data.