Measuring AI readiness in education: validation of the German AI readiness scale for educators

In recent years, artificial intelligence (AI) has become an integral part of various sectors, including education. In educational contexts, AI technologies offer the opportunity to support personalized, flexible, and inclusive learning experiences. They can assist teachers by providing feedback, analyzing student data, and enhancing instructional design through tools such as virtual simulators and intelligent tutoring systems and thus may improve the quality and effectiveness of teaching (Luckin and Holmes, 2016; Hwang et al., 2020; Gaur et al., 2024; Singh and Thakur, 2024). However, successfully and effectively integrating AI into teaching and learning environments requires that educators are adequately prepared to use these technologies (Qin et al., 2020). Research consistently points to limited AI readiness among teachers as a major barrier to effective implementation in schools (Zhang et al., 2023). While AI presents exciting opportunities, it also poses significant challenges (Hamilton et al., 2023). To harness its potential effectively, educators need not only technical proficiency in operating AI tools but also robust pedagogical knowledge for their meaningful integration into teaching and learning processes. This includes a comprehensive understanding of how AI functions, its educational potential, and the ability to critically evaluate its applications (Giannakos et al., 2024). Teachers must be capable of selecting appropriate AI tools aligned with specific instructional objectives and applying them in ways that genuinely enhance student learning (Ng et al., 2023). In addition, educators should be prepared to address the ethical dimensions of AI use – particularly with regard to data privacy, bias, and transparency (Giannakos et al., 2024). Or in the words of Wang et al. (2023) educators “are expected to develop an adequate understanding of AI and become educated users as well as educators” highlighting the dual responsibility to both competently apply and teach AI technologies. Assessing and strengthening teachers' AI readiness is therefore essential for ensuring that AI technologies are not only adopted but used effectively and responsibly in educational contexts. A lack of readiness risks not only limiting the pedagogical benefits of AI but also exacerbating existing digital inequalities. If educators are not adequately equipped, they may struggle to prepare students for life and work in an AI-driven world, thereby narrowing access to future technological and civic participation (Casal-Otero et al., 2023).

While AI readiness is often discussed in relation to organizations' capacity to adopt and implement AI technologies (e.g. Jöhnk et al., 2021; Holmström, 2022), it also applies at the individual level. The concept of AI readiness is especially important in the context of education, where educators need to transition from a basic understanding of AI to a more nuanced comprehension of how AI can support teaching and learning (Ayanwale et al., 2024). This is also reflected in the definition by Luckin et al. (2022) who refer to AI readiness as “a way to describe the transition that those working in education and their students need to make from not understanding what AI is and what AI can do, to being able to understand, in non-technical terms, what AI is capable of achieving”. This shift is particularly significant because AI in education not only involves using digital tools but also reimagining how educational processes can be enhanced through intelligent technologies (Walia and Kumar, 2022).

AI readiness is a multifaceted construct that encompasses various dimensions. According to Karaca et al. (2021), AI readiness can be broken down into four core components: cognitive readiness, the ability to use AI effectively, the vision for AI's role in education, and ethical considerations related to its application. Each of these components plays a critical role in determining how well educators can integrate AI into their teaching practices and curricula. Cognitive readiness involves understanding the theoretical concepts behind AI, while the ability to use AI refers to practical skills in applying AI tools. The vision for AI concerns how educators perceive AI's potential to reshape educational outcomes, and ethics addresses the concerns surrounding privacy, fairness, and transparency in AI applications.

AI literacy is widely defined as a set of competencies – including technical knowledge, ethical awareness, critical evaluation, and practical application – enabling individuals to understand, interact with, and responsibly use AI technologies (Long and Magerko, 2020; Ng et al., 2021; Allen and Kendeou, 2024; Almatrafi et al., 2024). Current studies show that AI literacy is a critical enabler of AI readiness, but the connection is nuanced: AI literacy alone does not guarantee readiness for the use of AI in education, as factors such as confidence, perceived relevance, self-efficacy, and contextual support also play significant roles (Zhao et al., 2022; Wang et al., 2023; Du et al., 2024; Özüdoğru and Durak, 2025). Hence, AI literacy is a necessary but not sufficient condition for AI readiness; interventions should address mediating factors to fully realize the benefits of AI literacy in educational practice.

Despite its growing importance, there is a lack of validated tools for assessing AI readiness, particularly in educational settings. Existing scales, such as the one developed by Karaca et al. (2021) for medical students, have focused on specific domains. While there have been some efforts to extend AI readiness scales to the educational context in China (Wang et al., 2023) and Turkey (Ramazanoglu and Akın, 2025), a comprehensive tool tailored to educators in German-speaking countries remains underexplored. This gap poses a significant challenge for both research and practice: without reliable and context-sensitive instruments, it becomes difficult to compare findings across studies, monitor progress over time, or design targeted professional development programs aimed at fostering AI readiness among educators. Moreover, the absence of such tools limits the ability of policymakers to make evidence-based decisions regarding curriculum development and teacher training standards. Addressing this gap is therefore essential for supporting educators in adapting to the evolving demands of AI-enhanced learning environments.

This study aims to address this gap by validating the German version of an AI readiness scale for educators, specifically focusing on its reliability and construct validity. The scale is designed to measure the four key dimensions of AI readiness: cognition, ability, vision, and ethics. Initially developed by Karaca et al. (2021) for assessing AI readiness in medical students, the scale has been adapted for teachers and translated into Chinese by Wang et al. (2023). This study builds upon those efforts by assessing the psychometric properties of the scale in a German-speaking context, ensuring its applicability and reliability in this specific educational setting.

Additionally, this study explores the correlations between AI readiness and key variables such as digital competencies, age, gender, and professional experience. These relationships provide insights into possible factors associated with educators' adoption of AI and can inform the design of professional development programs tailored to enhance AI readiness (Yue et al., 2024). Moreover, identifying such patterns may contribute to a better understanding of the conditions under which AI integration is most likely to succeed in educational settings.

1. What are the psychometric properties of the German version of the AI readiness scale for educators, and how well does it measure the different dimensions of AI readiness (cognition, ability, vision, and ethics)?

2. To what extent do digital competencies correlate with educators' AI readiness?

3. Are there significant differences in AI readiness based on demographic variables such as age, gender, and professional experience?

By addressing these questions, the study aims to contribute to a deeper understanding of the underlying aspects associated with AI readiness in education, providing valuable insights for both researchers and practitioners in the field of educational technology.

The data used to answer the research questions came from a supplementary survey conducted within a swissuniversities project, carried out in German-speaking Switzerland from 2021 to 2025. Out of 188 teachers who completed the main study, 153 (81.38%) answered the supplementary survey about AI. This study is based solely on data from these 153 participants.

The sample consisted of N = 153 teachers who were responsible for supervising and guiding student teachers in the context of their professional development. The participants were predominantly from primary and secondary schools, with a broad distribution of subjects taught. Participants' ages ranged from 24 to 63 years, with a mean age of M = 41.23 years (Mdn = 40 years). The sample was diverse in terms of gender, with 60.8% identifying as female and 38.6% as male and 0.7% did not indicate their gender. The sample also included a range of professional experience, with 8.5% of participants having fewer than 5 years of teaching experience, 44.4% having between 5 and 15 years, and 47.1% having more than 15 years of experience.

A convenience sampling approach was employed, with invitations sent to educators from various educational institutions in the region. Data were collected through an online survey, which was distributed via email. The survey included several sections, such as the supplementary questionnaire featuring the newly developed AI Readiness Scale, measures of digital competencies, and demographic questions (including age, gender, and years of teaching experience). Completion of the survey of the current study took approximately 10 min, and participation was entirely voluntary.

The study adhered to established ethical standards for research involving human participants. The Ethics Committee of the institution (St.Gallen University of Teacher Education) confirmed that the project complies with the principles of ethically sound research practice and scientific integrity as outlined in the Code of Scientific Integrity of the Swiss Academies of Arts and Sciences. Informed consent was obtained from all participants, and all responses were anonymized to ensure confidentiality. Participants were also informed of their right to withdraw from the study at any time without any negative consequences.

The primary tool used in this study was the newly translated German version of the AI Readiness Scale for Educators. The scale was originally constructed for medical students by Karaca et al. (2021) and later adapted for teachers and translated into Chinese by Wang et al. (2023). This scale was developed through a multi-step process, ensuring both its linguistic and conceptual validity. First, the original scale by Wang et al. (2023) underwent a rigorous translation process from English to German, followed by back-translation to ensure the accuracy and consistency of the items across languages. Furthermore, a native speaker checked the translation for clarity and language quality. Experts in both AI and education reviewed the translated scale to ensure that the content was relevant, clear, and culturally appropriate for German-speaking educators. During this adaptation, some items were slightly reworded to reflect the educational context and teaching practices in German-speaking regions, while preserving the original constructs and number of items. This process resulted in a refined version of the AI Readiness Scale that was both linguistically and contextually tailored to the educational context in German-speaking regions. The scale consists of 18 items and measures four key dimensions of AI readiness: cognitive readiness (i.e. Cognition), ability to utilize AI (i.e. Ability), vision for AI in education (i.e. Vision), and ethical considerations (i.e. Ethics). Participants rated their agreement with each item on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). A description of each dimension and respective sample items in English are depicted in Table 1. The full translated German version of the scale can be retrieved from Table A1 in the  Appendix.

Digital competencies: To assess participants' digital competencies, we used the subscales “Digital Skills” and “Perceived Usefulness of Digital Media” from the Fit4digit questionnaire (Zahn et al., 2019). The “Digital Skills” subscale consisted of 6 items (e.g. “I have adequate resources to use digital media in my work”; α = 0.81). The “Perceived Usefulness of Digital Media” subscale included 4 items (e.g. “Digital media are helpful for my work”; α = 0.88). Participants rated their agreement with each item on a Likert scale ranging from 1 (strongly disagree) to 5 (strongly agree). Although the Fit4digit questionnaire has not been explicitly validated for the teachers, it has been validated in German-speaking educational and work-related contexts. In the present sample, internal consistency was acceptable for both subscales, suggesting that the measures functioned reliably for teachers.

Demographic information: Participants were asked to provide their age in years, gender (male, female, other), and the number of years they had been teaching (i.e. teaching experience).

Data were analyzed using SPSS (Version 28) (IBM Corp, 2021) and Mplus (Version 8.4) (Muthén and Muthén, 2019) for statistical modeling. Descriptive statistics, such as means, standard deviations, and frequencies, were computed to summarize the demographic characteristics of the participants.

To assess the psychometric properties of the AI Readiness Scale for Educators, confirmatory factor analysis (CFA) was performed in Mplus, evaluating model fit indices such as the ratio of chi-square to the degrees of freedom (χ²/df), the root mean square error of approximation (RMSEA), the comparative fit index (CFI), the Tucker-Lewis Index (TLI), and the standard root mean square residual (SRMR). The criterion χ²/df was adopted instead of χ² because χ² is excessively sensitive to the sample size (Hu and Bentler, 1999). Acceptable fit between the developed model and the collected data is indicated by a χ²/df value of ≤5 (Iacobucci, 2010), a CFI and TLI of >0.90 (Russell, 2002), and an SRMR of <0.08 (Hu and Bentler, 1999). An RMSEA of between 0.08 to 0.10 provides a mediocre fit and below 0.08 shows a good fit (MacCallum et al., 1996). Internal consistency was assessed using Cronbach's alpha for each dimension (i.e. each construct of the scale). To empirically evaluate the stability of the CFA model with N = 153, a Monte-Carlo simulation was conducted in Mplus using the population model parameters from the study.

To evaluate whether the four subscales of the AI Readiness Scale measure distinct constructs, the heterotrait-monotrait ratio of correlations (HTMT) was calculated.

Measurement invariance across gender (male vs female) and age (younger vs older, median split groups) was examined using multiple-group CFA in Mplus. We tested configural, metric, and scalar invariance following (Cheung and Rensvold, 2002) using changes in CFI (ΔCFI ≤ 0.01) as the primary criterion for invariance.

To examine the relationships between AI readiness and the key teacher characteristics (i.e. digital competencies, age, gender, and teaching experience), Pearson correlation coefficients were computed.

Confirmatory factor analysis (CFA) was conducted to validate the four-factor structure of AI readiness. The results are graphically depicted in Figure 1 and confirmed that the four constructs - Cognition, Ability, Vision, and Ethics - constitute a valid model for measuring AI readiness in educators (χ²/df = 2.19, p < 0.001, CFI = 0.92, TLI = 0.903, RMSEA = 0.088, 90% CI [0.074, 0.102], SRMR = 0.054). All fit indices fall within a range between acceptable to good fit. Factor loadings of the items ranged between 0.601 and 0.880 and were statistically significant for all items (p < 0.001).

The internal consistency of the four constructs was assessed using Cronbach's alpha. The Cognition subscale demonstrated excellent reliability (α = 0.865), and the Ability subscale showed very high reliability (α = 0.921), indicating strong internal consistency. The Vision (α = 0.789) and Ethics (α = 0.797) subscales demonstrated good reliability, which is acceptable given their smaller number of items. Corrected item-total correlations for all constructs ranged from moderate to strong, and “alpha if item deleted” values indicated that no items substantially reduced scale reliability. These values are presented in Table 2.

Although not all replications were successful, 23 Monte-Carlo simulations confirmed that factor loadings and covariances were reliably recovered. Across these simulations, factor loadings ranged from 0.788 to 1.079, latent covariances from 0.466 to 0.645, and residual variances from 0.216 to 1.053. Average fit indices indicated excellent model fit, with a mean RMSEA of 0.020 (SD = 0.014) and a mean CFI of 0.993 (SD = 0.008), providing empirical support for the stability of the CFA results given the moderate sample size. Overall, these findings indicate that the scale items are coherent and consistently measure their respective constructs.

HTMT values between latent factors ranged from 0.732 to 0.921. All factor pairs except Cognition - Ability (HTMT = 0.921) fell below the commonly recommended threshold of 0.85, indicating generally good discriminant validity. The slightly higher value for Cognition -Ability suggests a moderate overlap between these two dimensions, which is theoretically plausible given their conceptual relatedness.

Multiple-group CFA testing configural, metric, and scalar invariance for gender showed that across all three models, global fit indices remained identical (CFI = 0.887, RMSEA = 0.101), and χ² differences across models were negligible. This pattern indicates that the imposed equality constraints did not meaningfully affect model fit. According to the recommended criterion of ΔCFI ≤ 0.01 (Cheung and Rensvold, 2002), these results support measurement invariance across gender. Similarly, for age groups, configural, metric, and scalar invariance models yielded identical RMSEA (0.093) and CFI (0.903), demonstrating measurement invariance between younger and older teachers. These results suggest that the AI Readiness Scale measures the same constructs consistently regardless of gender or age, supporting meaningful comparisons across these subgroups.

The study further explored the relationships between AI readiness and key teacher characteristics, including digital competencies, and demographic variables such as age, gender, and teaching experience. Descriptive statistics are depicted in Table 3 and correlations are displayed in Table 4. The findings revealed several significant correlations:

Digital competencies: Teachers with higher digital competencies (digital skills and perceived usefulness of digital media) demonstrated greater AI Readiness across all four dimensions, as well as on the overall scale (r = 0.65, p < 0.001). This suggests that teachers who are more digitally proficient and perceive digital media as useful are better prepared to adopt AI technologies in their teaching practices.

Gender: Male teachers were found to have higher AI readiness scores (total scale) compared to female teachers (r = 0.16, p < 0.05), suggesting that gender may influence teachers' perceptions and attitudes toward AI adoption in education. With regard to the factors, male teachers scored slightly higher on ability (M_male = 2.68, SD = 0.97; M_female = 2.47, SD = 1.00) and vision (M_male = 3.21, SD = 0.85; M_female = 3.00, SD = 0.90).

Teaching experience and age: Teachers with more teaching experience exhibited lower AI readiness (total scale; r = −0.18, p < 0.05). With regard to the factors, only vision showed no significant correlation with teaching experience. For age, we found that older teachers exhibited lower AI readiness regarding the factor cognition (r = −0.17, p < 0.05), indicating potential challenges in AI-related knowledge acquisition or conceptual understanding among more senior educators.

Controlling for digital competencies: When digital competencies were included as a control variable, the previously observed correlations between AI readiness and both gender as well as teaching experience and age were no longer statistically significant. This finding underscores the central mediating role of digital competencies in AI readiness. In other words, differences in AI readiness attributed to gender, age, or teaching experience may largely reflect underlying disparities in digital skill levels rather than demographic effects.

This study aimed to validate the German version of the AI Readiness Scale for Educators and to explore how various teacher characteristics - such as digital competencies, age, gender, and teaching experience - are associated with AI Readiness. The findings suggest that AI readiness is a multidimensional construct, with distinct but interrelated factors that shape educators' readiness to adopt AI in the classroom. Specifically, the results highlight the importance of digital competencies in determining how teachers perceive and engage with AI technologies.

The results of the Confirmatory Factor Analysis (CFA) confirm that the proposed four-factor model of AI Readiness, which includes Cognition, Ability, Vision, and Ethics, is a valid and reliable measure for assessing AI readiness in teachers. The good model fit indices and high factor loadings provide strong evidence that the scale accurately captures the key components of AI readiness, making it a useful tool for future research and practice. Discriminant validity analyses indicate that the four subscales generally measure distinct constructs, with only the Cognition - Ability pair showing a slightly higher overlap, which is theoretically plausible given their conceptual relatedness. Measurement invariance analyses further suggest that the scale assesses AI readiness consistently across gender and age groups. Importantly, the validation of this model was conducted on the German-language version of the AI Readiness Scale for Educators, broadening its applicability to educators in German-speaking regions and offering a reliable instrument for cross-cultural research and professional development programs.

One of the key findings of this study is the positive correlation between digital competencies and AI readiness. This highlights the critical role that digital skills play in educators' ability to integrate AI into their teaching practices. Teachers who possess higher digital skills and perceive digital media as useful for their work tend to exhibit stronger AI readiness across all four factors. This finding is consistent with previous research, which has demonstrated that digital competencies are a key enabler for technology adoption in education (Antonietti et al., 2022). As AI technologies become increasingly integrated into educational contexts, it is essential for professional development programs to focus on enhancing teachers' digital competencies. The ability to effectively use digital tools not only facilitates the adoption of AI but also helps teachers better understand and engage with the broader implications of AI for teaching and learning.

Notably, when digital competencies were controlled for, previously significant associations between AI readiness and demographic variables such as gender, age, and teaching experience became non-significant. This underscores the central role of digital skills: rather than reflecting inherent demographic differences, disparities in AI readiness appear to stem largely from different levels of digital competencies. We will elaborate on this in the next paragraphs.

Initial analyses found that male teachers exhibited higher AI readiness compared to their female counterparts, particularly in the dimensions of Ability and Vision. However, this gender difference was no longer statistically significant once digital competencies were controlled for. This suggests that the gap may be driven more by differences in digital competences and confidence than by gender itself. Earlier studies have found that men often report greater confidence in using new technologies, possibly due to differences in exposure, self-efficacy, or social expectations (Gómez-Trigueros and Yáñez de Aldecoa, 2021; Guillén-Gámez and Rodríguez-Fernández, 2022). Our findings align with these broader societal trends, indicating that higher digital exposure and technological self-efficacy, rather than gender itself, shape AI readiness. Consequently, efforts to promote AI integration should focus on ensuring equitable access to digital skills development. Professional development programs, in particular, initiatives that support female educators in building digital skills and technological self-efficacy, may be especially effective in narrowing gender-related differences in AI readiness (Li, 2015).

The initially found negative correlations between teaching experience and AI readiness suggest that more experienced teachers are less likely to be prepared for AI adoption in the classroom. Interestingly, among the individual dimensions of AI readiness, only Vision was not significantly related to teaching experience. This suggests that experienced educators still recognize the potential and relevance of AI but may feel less equipped in terms of knowledge or practical ability to implement it. In terms of age, initial analyses revealed that older teachers showed lower scores specifically in the Cognition dimension, pointing to possible challenges in acquiring or understanding AI-related concepts. This finding aligns with previous research suggesting that older educators may be more resistant to adopting new technologies due to established pedagogical routines and limited exposure to emerging digital tools (Lucas et al., 2021).

However, as for gender, also these relationships lost statistical significance once digital competencies were controlled for, again underscoring the central role of digital skills in AI readiness. To address this, professional development efforts should prioritize not only AI-specific training but also the strengthening of basic digital competencies, particularly for senior educators. Initiatives such as mentoring programs, hands-on workshops, and peer learning opportunities could help build the necessary confidence and skills to engage with AI technologies (e.g. Hur, 2025; Schallert-Vallaster et al., 2025). As Lucas et al. (2021) emphasize, the integration of AI in education is not solely about learning to use new tools but also involves a shift in mindset and pedagogical practice. Supporting educators through this transition is key to ensuring inclusive and effective AI adoption across all levels of experience and age.

The findings of this study have important implications for the design of professional development programs aimed at fostering AI readiness among teachers. First, enhancing digital competencies should be a central focus of these programs, as teachers with higher digital skills are more likely to be prepared for AI adoption.

Furthermore, gender differences and the challenges faced by older and more experienced teachers highlight the need for targeted interventions. Given the rapid advancement of AI in education, programs should not only provide technical training but also consider the broader social and psychological factors that influence AI adoption. This includes the willingness to embrace change and to foster an innovative mindset that empowers educators to explore AI's potential in the classroom.

Despite the valuable insights this study provides, it is important to acknowledge both its strengths and limitations. A key strength is the validation of the German version of the AI Readiness Scale for Educators, offering a reliable and psychometrically sound tool for assessing AI readiness in German-speaking educational contexts. The strong model fit indices and high internal consistency confirm the robustness of the instrument. Additionally, by identifying significant correlations between AI readiness and demographic and individual factors such as digital competencies, gender, age, and teaching experience, the study contributes to a deeper understanding of the factors influencing educators' preparedness for AI integration. These findings provide useful insights for policymakers and educational institutions in designing targeted professional development programs.

However, some limitations must be acknowledged. The relatively small sample size (N = 153) of German-speaking teachers in Switzerland limits the generalizability of the findings to other educational contexts and falls below conservative rule-of-thumb thresholds for CFA (Wolf et al., 2013). Still, the stability of the CFA model was supported by both high factor loadings (0.601–0.880) and a Monte-Carlo simulation, which demonstrated reliable recovery of factor loadings and covariances, and excellent average model fit, indicating that the sample was adequate for validating the four-factor structure of the AI Readiness Scale. Additionally, the reliance on self-reported data introduces the potential for biases, such as social desirability bias and self-perception errors. Another limiting factor is that contextual factors such as school infrastructure, access to digital resources, or prior exposure to AI technologies were not explicitly assessed, which may have influenced the results. Future research could strengthen these findings by incorporating larger, more diverse samples, objective assessments of digital competencies, and relevant contextual variables. Furthermore, as this study is cross-sectional, it cannot establish causal relationships between AI readiness and influencing factors. Longitudinal studies would be beneficial in tracking changes over time and identifying causal mechanisms.

To build on these findings, future research should focus on expanding the study to include educators from diverse educational systems, allowing for cross-cultural comparisons of AI readiness. Investigating interventions designed to enhance AI readiness, such as targeted training programs, would provide actionable recommendations for improving professional development. Additionally, qualitative research, such as interviews and focus groups, could offer deeper insights into teachers' experiences, challenges, and needs regarding AI adoption. Finally, longitudinal studies tracking AI readiness over time could provide valuable information on how educators adapt to AI technologies and what strategies are most effective in supporting their development.

This study successfully validated the German version of the AI Readiness Scale for Educators, confirming its reliability and validity as an effective tool for assessing educators' preparedness to integrate AI into their teaching practices. Further, the findings underscore the critical role of digital competencies in fostering AI readiness: After controlling for digital competencies, previously observed differences related to gender, age, and teaching experience were no longer statistically significant. This highlights the importance of targeted professional development programs that focus on strengthening digital competences together with AI-specific affordances across all educator groups - regardless of demographic background. Moving forward, continued research and the implementation of strategic measures can help ensure that all educators develop the competencies necessary to effectively integrate AI into their pedagogical practices, thereby advancing AI adoption across educational contexts.

The authors would like to thank the teachers for their participation in the study. Special thanks go to the project team of the project “Online-based self-learning courses for mentor teachers”, that is, Corinne Wyss, FHNW School of Education, Samuel Krattenmacher, Schaffhausen University of Teacher Education, Kerstin Bäuerlein, FHNW School of Education, Nina Hüsler, FHNW School of Education, Eva Weingartner, St. Gallen University of Teacher Education, Alexandra Roggensinger, St. Gallen University of Teacher Education, and Angela Jochum, St. Gallen University of Teacher Education.