Background/Objectives: Artificial intelligence (AI) is transforming higher education (HE), reshaping teaching, learning, and feedback processes. Feedback generated by large language models (LLMs) has shown potential for enhancing student learning outcomes. However, few empirical studies have directly compared the quality of LLM feedback with feedback from novices and experts. This study investigates (1) the types of prompts needed to ensure high-quality LLM feedback in teacher education and (2) how feedback from novices, experts, and LLMs compares in terms of quality. Methods: To address these questions, we developed a theory-driven manual to evaluate prompt quality and designed three prompts of varying quality. Feedback generated by ChatGPT-4 was assessed alongside feedback from novices and experts, who were provided with the highest-quality prompt. Results: Our findings reveal that only the best prompt consistently produced high-quality feedback. Additionally, LLM feedback outperformed novice feedback and, in the categories explanation, questions, and specificity, even surpassed expert feedback in quality while being generated more quickly. Conclusions: These results suggest that LLMs, when guided by well-crafted prompts, can serve as high-quality and efficient alternatives to expert feedback. The findings underscore the importance of prompt quality and emphasize the need for prompt design guidelines to maximize the potential of LLMs in teacher education.

Artificial intelligence (AI) in higher education (HE) is reshaping teaching and learning, and feedback provided by large language models (LLMs) seems to have an impact on student learning. However, few empirical studies have compared the quality of LLM feedback with the feedback quality of real persons. Therefore, this study addresses the following questions: What prompts are needed to ensure high-quality LLM feedback in HE? How does feedback from novices, experts, and LLMs differ in terms of quality and content accuracy? We developed a learning goal with three errors and a theory-based manual to evaluate prompt quality. Specifically, three prompts of varying quality were created and used to generate feedback from ChatGPT-4. We provided the highest-quality prompt to novices and experts. Our results showed that only the best prompt produced consistently high-quality feedback. Additionally, LLM and expert feedback were significantly better than novice feedback, with LLM feedback being both faster and better than expert feedback in the categories of explanation, questions, and specificity. This suggests that LLM feedback can be a high-quality and efficient alternative to expert feedback. However, we postulate that prompt quality is crucial, highlighting the need for prompting guidelines and human expertise.

The use of artificial intelligence in teacher education is seen as having great potential. AI tools can help relieve teachers’ workloads and support learners individually in the classroom. However, they are still relatively rarely used by practicing teachers and currently, there are nearly no empirical findings regarding the use of AI tools as well as the factors influencing AI usage behavior among pre-service teachers. Using the Theory of Planned Behavior, the following article examines which factors can explain the weekly use of AI by pre-service teachers. In an empirical-quantitative study, n=931 pre-service teachers across Germany were asked about their perceived usefulness, their perceived subjective norm, their self-efficacy beliefs and their behavioral intention to use AI tools in their future work as teachers in schools. Pre-service teachers also provided information about their average weekly AI usage time. The results of a structural equation model show that the behavioral intention to use AI tools in the future can be highly significantly explained by the perceived usefulness of AI tools, the subjective norm of AI use, and the participants’ self-efficacy beliefs. In addition, pre-service teachers’ behavioral intention and self-efficacy beliefs are significant predictors of the weekly duration of AI use. Specifically, the correlation between pre-service teachers’ self-efficacy beliefs and their weekly AI use is mediated by their behavioral intention. The results of the study thus provide important insights into how to promote the use of AI tools in teacher education.

The release of ChatGPT in November 2022 has sparked discussions about integrating Generative Artificial Intelligence (GenAI) into teacher education. Teacher educators, as key facilitators of pre-service teachers, play a critical role in determining the successful adoption of GenAI. Their perceptions guide curriculum redesign, define best practices, and shape pre-service teachers’ experiences with GenAI, potentially multiplying its impact across future classrooms. This mixed-methods study explored Danish teacher educators’ perceptions of GenAI (n = 91), focusing on its transformative potential, challenges, and the knowledge pre-service teachers need to acquire. The study employed innovative methods, including GenAI-enhanced thematic analysis and Natural Language Processing, to analyze qualitative and quantitative data. Results highlight diverse perceptions, ranging from enthusiasm for fostering innovative teaching practices (AI didactics) to concerns about ethical implications, assessment challenges (AI assessment), and safeguarding basic skill acquisition. Teacher educators emphasized the need for pre-service teachers to acquire AI literacies—AI as a teaching tool, content, and learning tool—framed within ethical, cultural, and democratic boundaries. Mediation analyses revealed that intrinsic motivation and confidence significantly predicted teacher educators’ use of GenAI, fully mediating their perceptions of its potential. Findings highlight the importance of professional development initiatives that offer hands-on GenAI experiences and informal, collaborative opportunities to address its rapid development. This study highlights GenAI’s transformative potential for teacher education and the complex skillset pre-service teachers need to engage critically and responsibly, ensuring its integration aligns with educational and societal values.

Virtual reality (VR) in teacher education offers an immersive environment for microteaching which, when coupled with peer-feedback can enhance pre-service teachers’ competence. This study employed a mixed-method quasi-experimental design to compare VR-based microteaching with real-world microteaching among 150 pre-service teachers ( n = 73 feedback providers, n = 77 feedback receivers). Quantitative analyses revealed similar perceptions of feedback across both conditions, but minor differences in feedback quality, particularly in the specificity of comments. Qualitative findings highlighted VR’s potential for safe, standardized practice and guided reflection, while also pointing to limitations in simulating complex classroom interactions. The study highlights VR’s potential for feedback-oriented teacher training and its potential to support the development of professional teaching competencies.

Feedback plays a central role in learning, yet pre-service teachers' engagement with feedback depends not only on its quality but also on their perception of the feedback content and source. Large Language Models (LLMs) are increasingly used to provide educational feedback; however, negative perceptions may limit their practical use, and little is known about how pre-service teachers' perceptions and behavioral responses differ by feedback source. This study investigates how the perceived source of feedback - LLM, expert, or peer - influences feedback perception and uptake, and whether recognition accuracy and feedback quality moderate these effects. In a randomized experiment with 273 pre-service teachers, participants received written feedback on a mathematics learning goal, identified its source, rated feedback perceptions across five dimensions (fairness, usefulness, acceptance, willingness to improve, positive and negative affect), and revised the learning goal according to the feedback (i.e. feedback uptake). Results revealed that LLM-generated feedback received the highest ratings in fairness and usefulness, leading to the highest uptake (52%). Recognition accuracy significantly moderated the effect of feedback source on perception, with particularly positive evaluations when LLM feedback was falsely ascribed to experts. Higher-quality feedback was consistently assigned to experts, indicating an expertise heuristic in source judgments. Regression analysis showed that only feedback quality significantly predicted feedback uptake. Findings highlight the need to address source-related biases and promote feedback and AI literacy in teacher education.