A Comparative Analysis of Pre-service Teachers’ Readiness for AI Integration

Journal Pre-proof A Comparative Analysis of Pre-service Teachers’ Readiness for AI Integration Margarida Lucas, Pedro Bem-haja, Yidi Zhang, Carmen Llorente-Cejudo, Antonio Palacios-Rodríguez PII: S2666-920X(25)00036-0 DOI: https://doi.org/10.1016/j.caeai.2025.100396 Reference: CAEAI 100396 To appear in: Computers and Education: Artificial Intelligence Received Date: 6 August 2024 Revised Date: 22 January 2025 Accepted Date: 14 March 2025 Please cite this article as: Lucas M., Bem-haja P., Zhang Y., Llorente-Cejudo C. & Palacios-Rodríguez A., A Comparative Analysis of Pre-service Teachers’ Readiness for AI Integration, Computers and Education: Artificial Intelligence, https://doi.org/10.1016/j.caeai.2025.100396. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2025 Published by Elsevier Ltd. A Comparative Analysis of Pre-service Teachers’ Readiness for AI Integration Addresses and affiliations: Margarida Lucas (Corresponding Author) University of Aveiro, Portugal, Laboratory of Digital Contents, The Research Centre on Didactics and Technology in the Education of Trainers (CIDTFF), Department of Education and Psychology E-mail: [email protected] Campus de Santiago 3810-193 Aveiro, Portugal ORCID: https://orcid.org/0000-0002-7438-5287 Pedro Bem-haja University of Aveiro, Portugal, Center for Health Technology and Services Research (CINTESIS), Department of Education and Psychology E-mail: [email protected] ORCID: https://orcid.org/0000-0002-7547-5743 Yidi Zhang University of Aveiro, Portugal, Laboratory of Digital Contents, The Research Centre on Didactics and Technology in the Education of Trainers (CIDTFF), Department of Education and Psychology E-mail: [email protected] ORCID: https://orcid.org/0000-0001-5557-0906 Carmen Llorente-Cejudo University of Sevilla, Spain, Department of Didactics and Educational Organization, Faculty of Education Sciences [email protected] ORCID: https://orcid.org/0000-0002-4281-928X Antonio Palacios-Rodríguez University of Sevilla, Spain, Department of Didactics and Educational Organization, Faculty of Education Sciences E-mail: [email protected] ORCID: https://orcid.org/0000-0002-0689-6317 Acknolwedgements: This work is financially supported by National Funds through FCT – Fundação para a Ciência e a Tecnologia, I.P., under the project 2021.03379.CEECIND, 2022.14365.BD and UIDB/00194/2020. Conflicts of interest: The authors have no conflict of interest to declare. Journal Pre-proof A Comparative Analysis of Portuguese and Spanish Pre-service Teachers’ Readiness for AI Integration Abstract: Artificial Intelligence (AI) is impacting the way we teach and learn. However, there are still substantial barriers to the integration of AI in educational settings, as teachers often lack the necessary competences to understand and use AI for pedagogical purposes. Therefore, in the early stages of teacher preparation, it is essential to assess the readiness of pre-service teachers for this transition in different initial teacher education (ITE) contexts. This study aims to investigate the readiness of pre-service teachers in Portugal and Spain to integrate AI into future teaching practices. To this end, a study was conducted with 203 preservice teachers from both countries that compared results regarding different dimensions: trust in AI, knowledge of AI, digital competence and ITE training for AI. The study further examined the relations between these dimensions. The results showed no significant differences between the two countries for trust in AI, knowledge of AI and digital competence. Significant differences were observed for ITE training for AI, with the Spanish sample scoring higher. However, these differences were limited to the distinction between disagreement and neutrality, suggesting that pre-service teachers in both countries are not fully ready to integrate AI into their future teaching practices. The results also showed that ITE training for AI had almost no significant influence on both samples’ trust in AI, knowledge of AI and digital competence, suggesting that ITE in both countries are not integrating AI into their programmes. The study identifies aspects within ITE programmes that need to be reviewed to adequately qualify future teachers for AI readiness and provides practitioners, ITE providers and policy makers with insights to consider when updating ITE programmes. Keywords: Teacher education, Artificial Intelligence, Digital Competence, Preparedness, Educational Technology. 1. Introduction Education, like many other areas of society, is facing profound challenges with the rise of artificial intelligence (AI). AI is being used in different educational settings, particularly in K-12 education (Holmes, 2023), but as its use continues to grow, the insufficient readiness of teachers may present an obstacle to the integration of AI in schools (Kim & Kwon, 2023; Zhang et al., 2023). Although there is some guidance on how to maximise the opportunities and manage the risks associated with AI and education (Miao & Holmes, 2023; Miao et al., 2021), many AI tools and what they entail are new to teachers, who often lack sufficient knowledge to integrate them for teaching purposes or to cultivate students’ AI literacy (Ng et al., 2023). If students fail to become AI literate, they risk exclusion from the evolving digital landscape (Casal-Otero et al., 2023). This is particularly important among preservice teachers, who need to develop a comprehensive understanding of AI, including its potential and associated risks, and become equipped with the necessary digital competence to effectively integrate it into their future pedagogical practices and promote their students’ digital competence (Casal-Otero et al., 2023; Chounta et al., 2022; Holmes, 2023; Ng et al., 2023). Digital competence is considered a prerequisite for knowledge of AI (Polak et al., 2022; Wang et al., 2023), while knowledge of AI is closely related to attitudes toward AI (Ayanwale, Frimpong, et al., 2024). Lucas et al. (2024) examined the relation between K-12 teachers’ trust in AI, their knowledge of AI, and their digital competence. The results indicated positive relations among all three variables, with knowledge of AI being a substantial predictor of trust in AI. Without knowledge of AI, the significant relation between digital competence and trust in AI ceases to exist. As such, pre-service teachers’ digital competence and knowledge of AI may not only be indicative of their ability to integrate technology confidently into their teaching but may also serve as a predictor of their readiness to adopt and use AI. In recent years, action plans for the development of K-12 teachers’ digital competence have been put into place, but less emphasis has been put on initial teacher education (ITE). This has led different researchers to stress the need to remodel ITE curricula (McGarr, 2024; Tomczyk, 2024). In Portugal, ITE has remained largely unchanged, with a low integration of digital technologies. Emphasis is placed on imparting technical knowledge, with limited understanding of teaching methods and the use of technology to support practices (Silva & Costa, 2022; Vieira & Pedro, 2023). In Spain, ITE has shifted in recent years. Digital competence is currently implemented in ITE, but it often remains superficial, focusing on instrumental use rather than on pedagogical and ethical uses. Digital competence is reported to be gained through informal experiences and self-teaching and not because of education programmes (Fraile et al., 2018; Novella-García & Cloquell-Lozano, 2021; Tárraga-Minguez et al., 2021). Overall, in both countries, Journal Pre-proof the level of digital competence is low among future teachers (Gallego-Arrufat et al., 2019), as well as among teachers at the beginning of their careers (Lucas & Bem-Haja, 2021). Against this competence gap, there is a pressing need to understand how ready pre-service teachers feel to integrate AI into their teaching practices and how well ITE is preparing them to do so. In this study, readiness is defined as the state of preparedness to effectively and confidently engage with AI in teaching and learning. It encompasses both personal factors (e.g. perceptions of knowledge and skills) and institutional factors (e.g. incorporation of AI into ITE). Consequently, the aim of this study is to compare two samples of pre-service teachers in terms of i) trust in AI, ii) knowledge of AI, iii) digital competence, and iv) ITE training for AI. The study also aims to examine the relations among them. A comparison of these variables in both countries and an understanding of the relationships among them could guide the development of a comprehensive approach to design teaching strategies that enhance digital competence and improve the integration of AI in ITE relevant to different key stakeholders, ranging from teachers to teacher educators to teacher education institutions to policymaking organizations. 2. Background 2.1 Digital competence and initial teacher training in Portugal and Spain The integration of digital competence into ITE is a complex process that requires the formulation of policies with a focus on the pedagogical use of digital technologies and the revision and alignment of curricula with existing digital competence frameworks. In both Portugal and Spain, the DigComp (Vuorikari et al., 2022) and DigCompEdu (Redecker, 2017) European frameworks for digital competence have been explicitly adopted as policy guiding tools for education and training (McGarr et al., 2021; Vieira & Pedro, 2023), but their presence and adoption in ITE is rather shy. In Portugal, formal policies for ITE do not include specific training in digital technologies or the development of digital competences in their curricula (Eurydice, 2023; Vieira & Pedro, 2023). An analysis of over 800 ITE curricula concludes that the integration of digital technologies is marked by processes of stagnation and even regression, with practices mainly focused on promoting technical knowledge in the use of devices, software or specific applications (Vieira & Pedro, 2023). Pre-service teachers express positive attitudes toward the pedagogical use of digital technologies, but perceive knowledge gained in classrooms to be only basic and general (Graça et al., 2021; Janeš et al., 2023). Research points at possible limiting factors: levels of digital innovation in Portuguese higher education institutions (HEI) are low and teachers’ digital competence, including that of teacher educators is insufficient (Mascarenhas et al., 2023; Vicente et al., 2020). In Spain, recent formal policies require that teacher-specific digital competences be included in ITE curricula as a mandatory element (Eurydice, 2023; McGarr et al., 2021). Existing research though, suggests that while training provided may equip teachers with the knowledge and skills to understand and use digital tools, it is insufficient to equip them for their future innovative pedagogical practices. Future teachers refer the need to be trained to integrate digital technologies into their teaching practices but acknowledge that the training received is insufficient (Fraile et al., 2018; Tárraga-Minguez et al., 2021; Villar et al., 2022). As in Portugal, possible reasons may relate with the digital competence of teacher educators, which could be improved (Cabero-Almenara et al., 2021), and the digital maturity of HEI, which is in their initial steps and should be assessed and validated (Fernández et al., 2023). Studies involving samples from the two countries are similar in the results they present and confirm findings previously described (Gallego-Arrufat et al., 2019; Graça et al., 2022). 2.2 Pre-service teachers’ readiness for AI integration into future teaching practice The readiness of pre-service teachers toward AI is a recent area of study, which helps to justify the limited number of studies conducted so far. Different authors stress the need for further studies (Janeš et al., 2023; Moorhouse, 2024), as the topic is rather absent in the context of teacher education (Sperling et al., 2024). Frimpong (2022) surveyed pre-service teachers in Ghana and found a positive impact of AI training on their interest and inclination to learn and promote AI content. Ayanwale, Frimpong, et al. (2024) further clarified that pre-service teachers’ engagement in AI learning in Ghana was influenced by factors such as attitude, anxiety, readiness, self-transcendence goals and confidence. Ojo (2023) implemented a professional development programme for pre-service computer science teachers in Nigeria, which resulted in a significant increase in their skills and abilities, but no increase in their perceived preparedness to use AI. Nevertheless, the perceived threat of AI among teachers decreased, suggesting a positive impact Journal Pre-proof of the training in building trust in AI learning. A limitation of these studies is that they assess direct outcomes before and after training, without considering the effectiveness of integrating AI literacy into curriculum planning or the influence of other factors, such as perceived knowledge of AI or digital competence. Zhang et al (2023) conducted a survey of pre-service teachers in Germany and identified perceived ease of use and usefulness as primary predictors of their willingness to use AI. Ayanwale, Adelana, et al. (2024) found a relation between knowledge of AI among pre-service teachers in Nigeria and their understanding and use of AI. However, no relation was found between knowledge of AI and emotional regulation, nor between the active use of AI. Moorhouse (2024) investigated the preparedness of initial language teachers in Hong Kong to use generative AI tools. The results showed that beginning teachers are generally prepared to use generative AI tools and recognise their potential to support their professional tasks, largely due to their experience with ChatGPT. However, beginning teachers are not adequately prepared to integrate them in teaching and have a limited understanding of them. While these studies provide valuable insights into the field, they do not provide comparative data that can substantiate the relation between different factors that may contribute for pre-service teachers’ readiness to incorporate AI into their future teaching practices. In addition, the majority relate to a single institution or country, and do not pertain to countries where digital competence policy documents are formally adopted for teacher education. In the specific contexts of Portugal and Spain, the readiness of pre-service teachers to integrate AI in future teaching has not been investigated yet. It remains unclear whether ongoing curriculum reforms and developments in digital competence are effectively incorporating AI knowledge into ITE, and whether pre-service teachers are benefiting from such initiatives. 2.3 Study aim As such, the aim of this study is to compare the results obtained by two samples regarding: i) trust in AI; ii) knowledge of AI; iii) digital competence; and iv) ITE training for AI. The study further examines the relations among them through the following research questions: RQ1: How do pre-service teachers in Portugal and Spain compare in their levels of trust in AI, knowledge of AI, digital competence and ITE training for AI? RQ2: What are the relationships between trust in AI, knowledge of AI, digital competence, and ITE training for AI among pre-service teachers in each country? 3. Material and methods 3.1 The questionnaire The questionnaire employed the instrument developed by Nazaretsky et al. (2022) to investigate teachers’ trust in AI. The instrument uses a five-point Likert scale and comprises 24 items categorized into three dimensions: Perceived benefits of AI in an educational setting (D1), Reasons for not trusting AI diagnosis (D2), i.e., anxieties related to using AI, AI lack of human characteristics and AI perceived lack of transparency, Working alongside AI to improve pedagogy (D3), i.e., self-efficacy, AI-based vs Human advice/recommendation, preferred means to increase Trust in AI, and required shift in pedagogy to adopt AI. The reliability coefficients for D1, D2, and D3 demonstrate a high level of reliability, with Cronbach’s alpha values exceeding 0.8 for each dimension for both Portuguese and Spanish samples (Robinson et al., 1991). The questionnaire comprised three additional sections: the first assessed participants’ knowledge of AI (KAI), the second their level of digital competence (DC), and the third assessed whether AI is incorporated into ITE training. KAI was measured using a six-point scale proposed by Chounta et al. (2022), with the following options: 1-I have never heard of AI; 2-Not sure what AI is; 3-I have limited knowledge about AI; 4-I know what AI is; 5-I know a lot about AI and 6-I am an expert in AI. DC was measured using a six-point scale inspired by Vuorikari et al. (2022). A label was applied to the first six proficiency levels of DigComp, ranging from A1 to C2, following the terminology used in other European frameworks (e.g. the Common European Framework of Reference for Languages), which are known to the students. ITE training for AI was measured using a five-point Likert scale and the following statements: “Curricular units explore the benefits and usefulness of AI in education” (BU), “Curricular units discuss ethical challenges related to the use of AI in education” (EC), and “I am being prepared to integrate AI into my teaching practice” (INT). 3.2 Sample and procedure Journal Pre-proof The study sample comprised 203 pre-service primary teachers from two mid-sized HEI: one in central Portugal (n=116; 81% female) another in the south of Spain (n=87; 78.2% female). The average age of Portuguese pre-service primary teachers was 23.8 and of Spanish one was 22.7. Pre-service teachers were students from courses taught by two authors. The purpose of the study and the link to the online questionnaire were shared by the two authors in their classes. Participation was entirely voluntary and anonymous, with informed consent obtained from all participants prior to their involvement. Participants were assured that their responses would be kept confidential and used solely for research purposes. Data collection occurred over a month, during which participants completed the questionnaires at their convenience. 3.3 Data analysis Data analysis was conducted using R and JASP. The distribution of variables was employed by the mean, median, 25th percentile, 75th percentile, and interquartile range (IQR). The Brunner-Munzel Test (Brunner & Munzel, 2000) was used to investigate the differences in the values of variables between the two samples. A bivariate correlation network analysis was used to investigate the relations among the variables. A partial correlation network analysis was conducted to determine whether these relations persisted when controlling for the effects of other variables. A relative importance analysis was conducted with the objective of determining the relative contribution of different variables. A relative importance analysis was conducted with the objective of determining the relative contribution of different variables in explaining a particular outcome. 4. Results 4.1 Pre-service teachers' TAI, KAI, DC, and ITE training for AI 4.1.1 Distribution of variables Figure 1 and Table 1 present the distribution of variables between Portugal and Spain. Figure 1. Distribution of variables between Portugal and Spain Note: D1: Perceived benefits of AI in an educational setting; D2: Reasons for not trusting AI diagnosis; D3: Working alongside AI to improve pedagogy; KAI: Knowledge of AI; DC: Digital competence; BU: Curricular units explore the benefits and usefulness of AI in education; EC: Curricular units discuss ethical challenges related to the use of AI in education; INT: I am being prepared to integrate AI into my teaching practice. Journal Pre-proof Table 1. Distribution of variables between Portugal and Spain D1 D2 D3 KAI DC BU EC INT Mean Portugal 3.69 3.53 3.24 3.45 3.21 2.71 2.33 2.05 Spain 3.79 3.45 3.23 3.62 3.46 3.02 2.90 3.14 Median Portugal 3.71 3.5 3.33 3 3 3 2 2 Spain 3.86 3.5 3.22 4 3 3 3 3 25th percentile Portugal 3.14 3.13 2.89 3 3 2 2 1.75 Spain 3.29 2.88 2.89 3 3 2 2 3 75th percentile Portugal 4.14 3.88 3.67 4 4 3 3 3 Spain 4.14 4 3.67 4 4 4 4 4 IQR Portugal 1 0.75 0.78 1 1 1 1 1.25 Spain 0.86 1.13 0.78 1 1 2 2 1 Note: D1: Perceived benefits of AI in an educational setting; D2: Reasons for not trusting AI diagnosis; D3: Working alongside AI to improve pedagogy; KAI: Knowledge of AI; DC: Digital competence; BU: Curricular units explore the benefits and usefulness of AI in education; EC: Curricular units discuss ethical challenges related to the use of AI in education; INT: I am being prepared to integrate AI into my teaching practice. Results reveal the mean values, median, 25th and 75th percentiles, and IQR exhibited close proximity across D1, D2, and D3. Both samples demonstrated similar central tendencies, with these variables falling within the range from neutral to agreement. However, a disparity in distribution was observed within KAI. Despite the 25th and 75th percentiles, and the IQR, remaining relatively close, the Portuguese sample presented an asymmetric distribution, with the median aligning entirely with the 25th percentile. In contrast, the Spanish sample displayed a higher mean and median. The KAI of Portuguese sample is more aligned with the third level, whereas the KAI of the Spanish sample is more closely aligned with the fourth level. Regarding DC, both samples exhibited similar mean and median values, as well as consistent 25th and 75th percentiles and IQR. This indicates a concurrence in central tendencies and distribution patterns, with both samples hovering at the intermediate B1 level. For BU, despite medians and 25th percentiles being identical, the higher mean, 75th percentile, and IQR in the Spanish sample indicate a greater shift and a more dispersed distribution. Consequently, while the overall BU of both samples is neutral, the Portuguese sample leans towards the disagreement to neutrality range, while the Spanish sample encompasses the disagreement to agreement range. For EC, the mean and median of the Portuguese sample were lower than those of the Spanish sample. While the 25th percentiles were identical, the higher IQR in the Spanish sample suggests a broader response range and greater variability, particularly within the upper limit range. In essence, the Portuguese sample reflects an overall disagreement, whereas the Spanish sample demonstrates a neutral stance, encompassing a range from disagreement to agreement. The most significant mean difference was observed for INT. This distinction was also reflected in the median and 25th percentile, with the Spanish sample displaying higher values. The greater variability indicated by the IQR in the Portuguese sample suggests a higher degree of dispersion below the median. Additionally, the Portuguese sample did not surpass neutrality, implying an overall disagreement, while the Spanish sample embodied a neutral stance. In summary, the descriptive statistics indicate a general concordance in D1, D2, D3, and DC. Despite the overall KAI of the two samples being similar, the Spanish sample exhibited a higher agreement. Regarding the three variables related to ITE training for AI (BU, EC and INT), the Spanish sample leaned towards a neutral position compared to the Portuguese sample but did not exhibit a strong consistency. 4.1.2 Comparison of variables in Portugal and Spain To verify the observed differences in the descriptive statistics, the Brunner-Munzel test was employed to compare the two samples (see Table 2). Table 2. Comparison of variables between Portugal and Spain Variables Brunner-Munzel Test Statistic df p P(Portugal < Spain) + ½P(Portugal = Spain) D1 0.944 186 .347 0.539 D2 -0.490 142 .625 0.479 D3 -0.353 176 .725 0.485 Journal Pre-proof KAI 1.625 190 .106 0.560 DC 1.713 187 .088 0.566 BU 1.927 169 .056 0.577 EC 3.015 134 .003 0.621 INT 9.859 138 < .001 0.798 Note1: D1: Perceived benefits of AI in an educational setting; D2: Reasons for not trusting AI diagnosis; D3: Working alongside AI to improve pedagogy; KAI: Knowledge of AI; DC: Digital competence; BU: Curricular units explore the benefits and usefulness of AI in education; EC: Curricular units discuss ethical challenges related to the use of AI in education; INT: I am being prepared to integrate AI into my teaching practice. Note2. Hₐ P(Portugal < Spain) + ½P(Portugal = Spain) ≠ ½ Results show there were no significant differences between the Portuguese and Spanish samples in D1 (BM=0.944, p=.347), D2 (BM=-0.490, p=.625), D3 (BM=-0.353, p=.725), KAI (BM=1.625, p=.106) and DC (BM=1.713, p=.088). However, there were significant or marginally significant differences in BU (BM=1.927, p=0.056), EC (BM=3.015, p=.003), and INT (BM=9.859, p<.001). When ties were evenly split, the probability of a random sample from Portugal being lower than a sample from Spain was found to be lower for BU (p = 0.577, 95% CI [0.498, 0.656]), for EC (p = 0.621, 95% CI [0.542, 0.700]), and for INT (p = 0.798, 95% CI [0.739, 0.858]), indicating a high probability of lower ITE training for AI among the Portuguese sample. Nevertheless, results also indicate that the overall tendency of BU, EC, INT in the Spanish sample is merely neutral. Consequently, the significant difference between the two samples is limited to the distinction between disagreement and neutrality. Therefore, this does not imply that the Spanish sample feels they are being prepared to integrate AI into their future teaching practices. 4.2 Relationships between TAI, KAI, DC and ITE training for AI 4.2.1 Network analysis A network analysis was conducted to evaluate the relations among all variables (see Figure 2). The results of the bivariate correlations show the Portuguese sample exhibited a network with 8 nodes and 9 edges, with a sparsity of 0.679. The Spanish sample exhibited a network with 8 nodes and 14 edges, with a sparsity of 0.500. When the two samples are combined, certain common relations emerge. For instance, D1 and D3 demonstrate a positive correlation, while D2 is negatively correlated with D3. KAI is positively correlated with DC, and BU, EC, INT are positively correlated with each other. When the two samples are compared, a weak negative relation is observed between D1 and DC for the Portuguese sample. The positive relations between KAI and BU and INT are pronounced. In the Spanish sample, the negative relation between D1 and D2 is evident. Additionally, KAI is positively correlated with D3. DC shows a positive relation with INT. Results also show that there are strong and positive relations between BU, EC, INT and D1, D2, D3. Journal Pre-proof Figure 2. Network analysis among all variables Note1: D1: Perceived benefits of AI in an educational setting; D2: Reasons for not trusting AI diagnosis; D3: Working alongside AI to improve pedagogy; KAI: Knowledge of AI; DC: Digital competence; BU: Curricular units explore the benefits and usefulness of AI in education; EC: Curricular units discuss ethical challenges related to the use of AI in education; INT: I am being prepared to integrate AI into my teaching practice. Note2: Blue lines indicate significant positive correlations, whereas red lines indicate significant negative correlations. The thickness of the lines represents the strength of the correlation. According to the results of the partial correlations, the Portuguese sample exhibited a network with 8 nodes and 5 edges, with a sparsity of 0.821 The Spanish sample exhibited a network with 8 nodes and 7 edges, with a sparsity of 0.750. When the two samples are combined, several common relations are maintained. For instance, D1 is positively correlated with D3, KAI is positively correlated with DC, and BU is positively correlated with EC and INT. Additionally, some relations that previously existed in both samples cease to exist. For example, KAI is no longer correlated with D1, D2, D3 and with BU, EC, INT. When the two samples are compared, INT is positively correlated with EC in the Portuguese sample. The previously observed negative relations between D1 and DC, as well as between D2 and D3 are not present. In the Spanish sample, D1 is positively correlated with INT, and D2 is negatively correlated with D3. The relations between D3 and BU, EC and INT cease to exist, but a new positive one emerges between DC and D2. In summary, the relations among the variables comprising trust in AI (D1, D2 and D3) are stable, as well as the relations among the ITE training for AI ones (BU, EC and INT). The interdependence between KAI and DC is observed, but neither KAI nor DC show a significant relation with the ITE training for AI variables in the two samples. In addition, BU, EC and INT, and D1, D2 and D3 are essentially one-to-one correspondences (BU to D1, EC to D2, INT to D3), but the relations between them almost disappear. This suggests that what pre-service teachers learn in the classroom does not have a significant effect on trust in AI, KAI, DC, in both samples. 4.2.2 Relative importance analysis To identify the observed results of the two network analysis and to determine the relative contribution of different variables for KAI, DC and INT, a relative importance analysis was conducted (see Table 3). We place particular emphasis on INT due to its prominence over other variables in descriptive, comparative and network analysis. Portugal Spain Bivariate correlations Partial correlations Journal Pre-proof Tomczyk, Ł. (2024). Digital competence among pre-service teachers: A global perspective on curriculum change as viewed by experts from 33 countries. Evaluation and Program Planning, 105. https://doi.org/10.1016/j.evalprogplan.2024.102449 UNESCO. (2022). K-12 AI curricula A mapping of government-endorsed AI curricula. https://unesdoc.unesco.org/ark:/48223/pf0000380602 Velander, J., Taiye, M. A., Otero, N., & Milrad, M. (2023). Artificial Intelligence in K-12 Education: eliciting and reflecting on Swedish teachers’ understanding of AI and its implications for teaching & learning. Education and Information Technologies. https://doi.org/10.1007/s10639-023-11990-4 Vicente, P. N., Lucas, M., Carlos, V., & Bem-Haja, P. (2020). Higher education in a material world: Constraints to digital innovation in Portuguese universities and polytechnic institutes. Education and Information Technologies, 25(6), 5815–5833. https://doi.org/10.1007/s10639-020-10258-5 Vieira, C. R., & Pedro, N. (2023). Weaknesses of ICT integration in the initial teacher education curriculum. Computers and Education Open, 5, 100150. https://doi.org/10.1016/j.caeo.2023.100150 Villar, L. B. E., Herrero, L. L., & Álvarez-López, G. (2022). UNESCO Strategy and Digital Policies for Teacher Training: The Deconstruction of Innovation in Spain. Journal of New Approaches in Educational Research, 11(1). https://doi.org/10.7821/NAER.2022.1.812 Vuorikari, R., Kluzer, S., & Punie, Y. (2022). DigComp 2.2: The Digital Competence framework for citizens. https://doi.org/10.2760/115376 Wang, B., Rau, P. L. P., & Yuan, T. (2023). Measuring user competence in using artificial intelligence: validity and reliability of artificial intelligence literacy scale. Behaviour and Information Technology, 42(9). https://doi.org/10.1080/0144929X.2022.2072768 Zhang, C., Schießl, J., Plößl, L., Hofmann, F., & Gläser-Zikuda, M. (2023). Acceptance of artificial intelligence among pre-service teachers: a multigroup analysis. International Journal of Educational Technology in Higher Education, 20(1). https://doi.org/10.1186/s41239-023-00420-7 Journal Pre-proof Highlights 1. Pre-service teachers’ readiness may present an obstacle to AI integration in schools 2. There is a pressing need to understand how ITE is preparing them to do so 3. Spanish pre-service teachers score slightly higher in ITE training for AI 4. Results show both samples are not fully ready to integrate AI into their future teaching 5. ITE providers and policymakers in Portugal and Spain should consider the revision of ITE programmes Journal Pre-proof Conflicts of interest: The authors have no conflict of interest to declare. Journal Pre-proof