Technology is advancing at an immense speed. As technology continued to advance, artificial intelligence (AI) also started to draw interest from education researchers, which gave rise to the field of AI in Education.  This field focuses on the application of AI to improve teaching and learning processes with the ultimate goal of transforming education (Crompton & Burke, 2023; Good, 1987) and enhance educational processes. AI tools can support personalized learning by adapting content to individual needs, offering real-time feedback to learners, and it can automate certain administrative and instructional tasks (Alexandrowicz, 2024).

In the context of science education, AI tools can assist teachers with modeling complex concepts, simulate experiments and support inquiry-based learning activities (Crompton & Burke, 2023). Several studies indicate that teachers report curiosity and enthusiasm towards AI integration (Alexandrowicz, 2024). Yet, studies show that their competence and confidence in using AI tools remains insufficient (Almuhanna, 2025). Moreover, research with preservice teachers also show similar results, indicating gaps in their awareness, literacy and content and technological knowledge related to AI (Ayanwale et al., 2024). As future educators, preservice teachers play a significant role in mediating between AI tools and learners and must be competent in this process.

Overall, it becomes evident that teachers, who are the immediate stakeholders in bringing AI into the classrooms, should be adequately prepared to utilize these tools effectively in their teaching and to navigate the challenges AI integration brings. Equally, preservice teachers, as the future teaching workforce, should also be sufficiently competent in applying educational technology in classrooms (Hu et al., 2025). Consequently, their preparedness for AI integration into education can be examined through the widely accepted Technological Pedagogical Content Knowledge (TPACK) framework.

TPACK framework is a commonly used model that describes the domains of knowledge teachers need for effectively integrating technology into their classrooms (Mishra & Koehler, 2006). While the TPACK framework is a valuable model for understanding teacher knowledge for technology integration (Celik, 2023), emerging technologies such as AI challenge its boundaries and creates a necessity for the revision of the model (Uyanik Aktulun et al., 2024).

AI-TPACK is an extended version of the TPACK framework. As Mishra and Koehler (2006) developed the original TPACK framework by building on Shulman’s (1986) PCK framework based on the necessity of integrating technology component into teacher knowledge, AI-TPACK builds on TPACK as the complexity of AI tools requires a more nuanced understanding of teacher knowledge. AI-TPACK offers this extended understanding, and it aims to measure teachers’ competencies in effectively integrating AI into their teaching. Additionally, it focuses on the interrelations between AI technology, pedagogical methods and content knowledge. This implies that as educators’ knowledge about AI technology improves, their existing knowledge also transforms accordingly. As the future of the educational workforce, it is critical for preservice teachers to be ready to integrate AI tools in their classrooms (Hu et al., 2025). Researchers deem AI literacy indispensable for future educators, since those with higher AI competence are expected to outperform their peers with lower AI competence (Ayanwale et al., 2024). Accordingly, recent studies examine readiness of preservice teachers to address these issues (Ayanwale et al., 2024; Karataş & Ataç, 2025). However, in the context of science education such studies are limited. Therefore, examining the AI-TPACK of preservice science teachers is essential.

Accordingly, this study addresses the following research question: What are preservice science teachers’ (PSTs) perceptions of AI-TPACK in science education?

This study adopts a quantitative research approach by utilizing survey research design. A total of 544 PSTs enrolled in science education programs in 6 universities participated in this study. Sample of this study consisted of second-, third- or fourth-year PSTs. First-year students were excluded from participation because they had not yet received sufficient theoretical understanding or experience related to the constructs measured in this study. A convenience sampling strategy was used, taking into account the availability and willingness of students to participate, as well as the cooperation of the institutions and relevant stakeholders. The participants were predominantly female, with 444 females (81.6%) and 100 males (18.4%). Preservice science teachers were distributed across years with 159 in their second year (29.2%), 215 in their third (39.5%) and 170 in the fourth (31.3%). The instrument used in this study to measure the PSTs’ AI-TPACK in science education was specifically adapted for this purpose from three existing instruments developed by MaKinster et al. (2010), Celik (2023) and An et al. (2023) after obtaining permissions from the authors of the original instruments. The participants’ AI-TPACK data were collected with the AI-TPACK Scale. All items were rated on a 5-point Likert scale (1 = strongly disagree, 5 = strongly agree), with higher scores indicating greater self-perceived competence in integrating AI into science education. The adapted AI-TPACK Scale used in the pilot study consisted of 84 items. EFA was conducted during the pilot study and CFA was conducted in the main study to evaluate construct validity, and an eight-factor structure was obtained. The final scale included 93 items across these eight factors: AI technological knowledge (AI-TK, 13 items), content knowledge (CK, 10), pedagogical knowledge (PK, 10), pedagogical content knowledge (PCK, 9), AI technological content knowledge (AI-TCK, 12), AI technological pedagogical knowledge (AI-TPK, 21), AI technological pedagogical content knowledge (AI-TPACK, 13) and AI ethics (AI-ETH, 5). Component scores were calculated as the means of their respective items, and no items were reverse-coded. The phrasing of the items was carefully adapted to the AI and science education context, with an emphasis on clarity and accuracy. Reliability coefficients ranged from acceptable to excellent across the eight components: AI-TK (α = .905), CK (α = .844), PK (α = .890), PCK (α = .886), AI-TCK (α = .894), AI-TPK (α = .936), AI-TPACK (α = .922) and AI-ETH (α = .753). The overall scale reliability was excellent (α = .974).

To address the research question, descriptive analysis was conducted for the eight components of the AI-TPACK framework. The participants reported moderately high mean scores across all components. The highest mean scores were found for the pedagogical knowledge (PK) (M = 3.99, SD = 0.47) and the pedagogical content knowledge (PCK) (M = 3.99, SD = 0.42). The lowest mean score was obtained in the AI ethics (AI-ETH) component (M = 3.51, SD = 0.57), followed by the AI technological pedagogical content knowledge (AI-TPACK) component (M = 3.71, SD = 0.52), which indicates slightly lower confidence in PSTs’ integrated competencies and ethical considerations. The descriptive findings revealed that PSTs reported relatively high levels PK and PCK domains which indicated their strong confidence in the pedagogical domains. This pattern was consistent with previous literature (Irmak & Yılmaz-Tüzün, 2019) The lower scores in AI-TPACK and AI-ETH may reflect the new and complex nature of integrating AI into science education. Moreover, the lower score in AI-ETH suggests that while PSTs may feel competent in other areas, their confidence in addressing ethical considerations related to AI integration is comparatively lower. This finding aligns with previous studies such as Karataş and Ataç (2025), which also reported that the ethics component had the lowest scores among teachers and preservice teachers. Overall, findings show that while PSTs generally perceive themselves as having relatively high AI-TPACK levels, there are still gaps in their competence, especially regarding the ethical dimension and the integration of AI across content and pedagogy. The results highlight the need for teacher education programs to address these areas, particularly ethical considerations and holistic AI integration, to ensure future science teachers are adequately prepared to use AI effectively and responsibly in their classrooms.

Alexandrowicz, V. (2024). Artificial intelligence integration in teacher education: Navigating benefits, challenges, and transformative pedagogy. Journal of Education and Learning, 13(6), 346. https://doi.org/10.5539/jel.v13n6p346 Almuhanna, M. A. (2025). Teachers’ perspectives of integrating AI-powered technologies in K-12 education for creating customized learning materials and resources. Education and Information Technologies, 30(8), 10343–10371. https://doi.org/10.1007/s10639-024-13257-y An, X., Chai, C. S., Li, Y., Zhou, Y., Shen, X., Zheng, C., & Chen, M. (2023). Modeling English teachers’ behavioral intention to use artificial intelligence in middle schools. Education and Information Technologies, 28(5), 5187–5208. https://doi.org/10.1007/s10639-022-11286-z Ayanwale, M. A., Adelana, O. P., Molefi, R. R., Adeeko, O., & Ishola, A. M. (2024). Examining artificial intelligence literacy among pre-service teachers for future classrooms. Computers and Education Open, 6, 100179. https://doi.org/10.1016/j.caeo.2024.100179 Celik, I. (2023). Towards Intelligent-TPACK: An empirical study on teachers’ professional knowledge to ethically integrate artificial intelligence (AI)-based tools into education. Computers in Human Behavior, 138, 107468. https://doi.org/10.1016/j.chb.2022.107468 Crompton, H., & Burke, D. (2023). Artificial intelligence in higher education: The state of the field. International Journal of Educational Technology in Higher Education, 20(1), 22. https://doi.org/10.1186/s41239-023-00392-8 Good, R. (1987). Artificial intelligence and science education. Journal of Research in Science Teaching, 24(4), 325–342. Hu, L., Wang, H., & Xin, Y. (2025). Factors influencing Chinese pre-service teachers’ adoption of generative AI in teaching: An empirical study based on UTAUT2 and PLS-SEM. Education and Information Technologies, 30(9), 12609–12631. https://doi.org/10.1007/s10639-025-13353-7 Irmak, M., & Yılmaz-Tüzün, Ö. (2019). Investigating pre-service science teachers’ perceived technological pedagogical content knowledge (TPACK) regarding genetics. Research in Science & Technological Education, 37(2), 127–146. https://doi.org/10.1080/02635143.2018.1466778 Karataş, F., & Ataç, B. A. (2025). When TPACK meets artificial intelligence: Analyzing TPACK and AI-TPACK components through structural equation modelling. Education and Information Technologies, 30(7), 8979–9004. https://doi.org/10.1007/s10639-024-13164-2 MaKinster, J. G., Boone, W., & Trautmann, N. M. (2010, March). Development of an instrument to assess science teachers’ perceived technological pedagogical content knowledge. National Association for Research in Science Teaching, Philadelphia, PA. Mishra, P., & Koehler, M. J. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. Teachers College Record, 108(6), 1017–1054. https://doi.org/10.1111/j.1467-9620.2006.00684.x Shulman, L. S. (1986). Those who understand: Knowledge growth in teaching. Educational Researcher, 15(2), 4. https://doi.org/10.2307/1175860 Uyanik Aktulun, O., Kasapoglu, K., & Aydogdu, B. (2024). Comparing Turkish pre-service STEM and Non-STEM teachers’ attitudes and anxiety toward artificial intelligence. Journal of Baltic Science Education, 23(5), 950–963. https://doi.org/10.33225/jbse/24.23.950