Human-AI co-creation significantly enhances college students' AI literacy through a structured pathway involving trust, technological readiness, AI explainability, and personalization features.
Objective: The main goal of this study was to investigate how human-AI co-creation behaviors impact college students' AI literacy development. Specifically, the researchers aimed to examine the factors that motivate college students to engage in co-creation with AI (including AI trust, Technology Readiness Index, AI explainability, and AI personalization) and to determine whether co-creation behaviors serve as mediators in enhancing AI literacy. The study sought to understand the transformation from traditional "use" of AI tools to collaborative "co-creation" relationships and their educational implications for preparing students for the AI era.
Methods: The researchers employed a quantitative survey methodology using Biggs' 3-P (Presage-Process-Product) learning model as the theoretical framework. Data were collected from 401 Chinese college students across various disciplines through an online survey platform "Credamo" and social media channels like Weibo between March and April 2024. The study utilized validated measurement scales adapted from previous research, including the Technology Readiness Index (TRI) scale from Parasuraman and Colby (2015), AI trust and explainability scales from Shin (2021), AI personalization scale from Gao et al. (2022), AI literacy scale from Wang et al. (2023), co-creation intention scale from Du et al. (2022), and co-creation behavior scale from Cherry (2014). All constructs were measured using 5-point Likert scales. The researchers conducted comprehensive reliability and validity tests, including convergent validity, discriminant validity, and composite reliability assessments. Structural equation modeling using IBM SPSS Amos v24 was employed to test the hypothesized relationships and mediating effects. Bootstrap methods were used to analyze the chain mediation model, establishing 95% confidence intervals to validate the indirect effects.
Key Findings: The study revealed significant positive relationships across all hypothesized pathways in the human-AI co-creation process. AI trust (β=0.273, p<0.001) and Technology Readiness Index (β=0.178, p<0.001) both significantly predicted students' co-creation intentions with AI. Similarly, AI explainability (β=0.258, p<0.001) and AI personalization (β=0.283, p<0.001) demonstrated strong positive effects on co-creation intentions. The results showed that co-creation intention significantly influenced co-creation behavior (β=0.669, p<0.001), which in turn substantially impacted AI literacy enhancement (β=0.557, p<0.001). Importantly, the study confirmed that co-creation intention and behavior serve as chain mediators in the relationship between individual characteristics (AI trust, TRI) and AI characteristics (explainability, personalization) with AI literacy outcomes. The mediation analysis revealed that collaborative behavior accounted for 59.8% of indirect effects, supporting the process-oriented nature of the 3-P model. Among the antecedent factors, AI personalization showed the strongest effect (β=0.283), followed by AI trust (β=0.273) and AI explainability (β=0.258), while TRI had a more moderate but still significant impact (β=0.178). The model demonstrated excellent fit indices (CMIN/DF=1.35, NFI=0.901, IFI=0.972, TLI=0.97, CFI=0.972, RMSEA=0.03), confirming the robustness of the theoretical framework.
Implications: These findings have profound implications for AI education and higher education policy in the artificial intelligence era. The study demonstrates that AI literacy can be effectively enhanced through informal co-creation experiences rather than solely through formal coursework, expanding educational approaches beyond traditional classroom instruction. The research provides empirical evidence that human-AI collaboration serves as a critical pathway for developing twenty-first-century digital competencies. For educators, the findings suggest that incorporating co-creation activities with AI into curricula can significantly improve students' AI literacy, technical skills, and preparedness for AI-integrated workplaces. The study highlights the importance of fostering students' trust in AI systems and technological readiness while ensuring AI tools provide adequate explainability and personalization features. From a policy perspective, the research supports investments in AI education infrastructure that emphasizes hands-on, collaborative experiences with AI technologies. The findings also suggest that higher education institutions should focus on developing students' positive attitudes toward AI collaboration rather than just technical skills. The study contributes to the growing body of literature on human-AI interaction in educational contexts and provides a validated framework for understanding how co-creation behaviors mediate the relationship between individual and technological factors in AI literacy development.
Limitations: The study acknowledges several important limitations that affect the generalizability and scope of the findings. First, the sample was concentrated exclusively among Chinese college students (66.3% female), limiting the cultural and geographical diversity necessary for broader generalizability across different educational systems and cultural contexts. The reliance on self-reported measures throughout the study introduces potential response bias, as participants might provide socially desirable answers or have difficulty accurately assessing their own AI literacy levels. The cross-sectional design prevents the establishment of true causal relationships and limits understanding of how co-creation behaviors and AI literacy develop over time. The study did not examine specific co-creation processes, strategies, or the quality of human-AI interactions, focusing instead on general behavioral intentions and outcomes. Additionally, the research was conducted during a specific time period (March-April 2024) when certain AI technologies and tools were available, which may not reflect the rapidly evolving AI landscape. The study also did not control for prior AI experience, educational backgrounds in technology, or variations in AI tool usage patterns among participants. The measurement instruments, while validated, were primarily adapted from existing scales that may not fully capture the nuanced aspects of human-AI co-creation in educational contexts.
Future Directions: The researchers suggest several promising avenues for future investigation to address current limitations and expand understanding of human-AI co-creation in education. Future studies should incorporate larger, more diverse samples across different countries, cultures, and educational systems to improve generalizability and cultural validity. Longitudinal research designs would provide valuable insights into how co-creation behaviors and AI literacy develop over time, allowing for the examination of sustained effects and long-term learning outcomes. Experimental methodologies could more accurately measure AI literacy enhancement and observe specific co-creation processes in controlled settings. Future research should explore additional factors influencing human-AI co-creation, such as individual differences, domain-specific knowledge, and contextual variables. Investigation of the AI perspective in co-creation processes could provide insights into how AI systems adapt and respond during collaborative interactions. Research could also examine specific types of AI tools and platforms to understand which features and capabilities most effectively support educational co-creation. Studies incorporating natural language processing techniques could analyze the quality of human-AI interactions, including communication patterns, error types, and their impact on learning outcomes. Future work should also explore the integration of human-AI co-creation into formal curricula, examining optimal pedagogical approaches and assessment methods. Additionally, research into ethical considerations, privacy concerns, and responsible AI use in educational co-creation contexts would provide important guidance for policy and practice.
Title and Authors: "Enhancing college students' AI literacy through human-AI co-creation: a quantitative study" by Haitao Wen, Xinyu Lin, Rongkang Liu, and Chang Su.
Published On: April 28, 2025
Published By: Interactive Learning Environments (Taylor & Francis Group)