AI Ed Wiki logoAI Ed Wiki
Use with AI

📄 Research Article

Science educators' AI literacy and AI usage in teaching: Implications for post-qualification programs

● high · 2026-08-16 · science-education, ai-literacy, teacher-education, professional-development, stem-education

Synthesis: Maurer, Brückner, Thyssen, Becker-Genschow, and Huwer (2026) surveyed n = 115 in-service secondary science educators from Germany and Switzerland (recruited via email to randomly selected schools; ages 25–64) using a cross-sectional mixed-methods design with three components: a validated 30-item AI Literacy competency test (nliteracy = 64 valid completions), a 15-item AI-usage questionnaire scored on five-point frequency scales (nusage = 81 after excluding 34 inconsistent cases, analysed at the subject level as nsubject = 115 datasets: 44 biology, 39 chemistry, 32 physics), and semi-structured follow-up interviews with n = 21 teachers. Teachers showed a solid but moderate general AI Literacy (M = 16.7 of 30, SD = 4.59), yet no significant correlation emerged between AI literacy and any aspect of AI use — including usage history (all p between .109 and 1.00) and frequency of perceived usage opportunities (p = .151 in class; p = .947 for preparation). AI was used most in the general "Information Search and Evaluation" area and least in subject-specific areas (Simulation, data processing, data acquisition), and only 43% of subject datasets reported current in-class use versus 63% for lesson preparation and follow-up. The authors conclude there is a clear need for AI-related post-qualification programs that focus on subject-specific AI literacy and concrete applications, proposing two TPACKAI development pathways.

Key Findings

Measures and instruments. AI literacy was assessed with the validated Hornberger et al. competency test (30 multiple-choice items spanning 16 sub-competency areas, including Ethics), collected online via LimeSurvey; usage was captured with a 15-item questionnaire using five-point Likert frequencies ("never" to "often"); and 21 semi-structured interviews were coded inductively by two coders (consensus coding). Statistical work used SPSS 30 with chi-square, Fisher's exact and Yates' corrections, Kruskal–Wallis, Mann–Whitney, Spearman, and eta coefficients at a 5% significance level; teachers were split by median into AI-literate and AI-illiterate groups.

AI literacy is moderate and unrelated to use. Mean literacy was 16.7/30 (SD 4.59), slightly below the Hornberger reference sample (n = 1286, M = 18.79, SD 5.57 on 31 items). AI literacy was independent of age, gender, and years of service (all p between .333 and .462) and showed no significant association with AI usage history or with how often teachers perceived opportunities to use AI.

Usage patterns and reasons. The most frequent in-class reason for using AI was "fun or interest" (n = 47), while the least frequent was AI being "part of the science discipline" (n = 17); for lesson preparation and follow-up, time saving (n = 63) and the importance of new technology (n = 62) dominated. Top reasons for non-use were "haven't dealt with it" (n = 34) and lack of free access to AI tools (n = 21).

Subject-specific gaps. Using the DiKoLAN framework's eight teaching areas, AI was used most in "Information Search and Evaluation" (n = 43 in class; n = 53 for preparation) and least in "Data Acquisition" (n = 7; n = 11). Subject-specific areas (SIM, DAP, DAQ) averaged only M = 15 mentions versus M = 28.8 for general areas, and only one subject correlation was significant — chemistry teachers citing "no subject-specific relevance" (Cramer's V = 0.437, pFisher < .001). Interviews confirmed teachers mainly use AI for student research and see adaptivity, material creation, and lesson planning as the most promising uses, while citing output quality and cognitive-decline concerns as obstacles.

Implications for post-qualification. The authors argue that general, subject-unspecific AI literacy (TKAI/TPKAI) does not drive classroom adoption and propose two pathways to develop TPACKAI: Path A from Pedagogical Content Knowledge toward using AI as a teaching tool, and Path B from AI content knowledge (TCKAI) toward teaching about AI as a subject. PD should be embedded in subject contexts rather than offered as standalone courses.

Limitations. Self-selection bias (AI-interested teachers may have over-responded), a small sample that may mask correlations, no AI definition provided to participants, and difficulty comparing literacy scores across different test versions limit generalizability.

Implication. Post-qualification for in-service science educators should build both general AI Literacy and practical, pedagogically grounded AI use, connecting to TPACK, Teacher Education, and Faculty Development within STEM Education.

Connected Concepts

Connected Articles

Citation

Maurer, N., Brückner, M., Thyssen, C., Becker-Genschow, S., & Huwer, J. (2026). Science educators' AI literacy and AI usage in teaching: Implications for post-qualification programs. Computers and Education Open, 100376. https://doi.org/10.1016/j.caeo.2026.100376