The KICK 4.0 research project focuses on the approach of effectively using large language model-based (LLM) generative AI technologies in laboratory-based engineering education. The aim of the project is to enable students and instructors to explore the benefits and limitations of generative AI systems in higher education.

Using a customized Design-Based Research (DBR) approach, an LLM-based AI tool is integrated into an ongoing course, evaluated iteratively, and refined. To this end, criteria for determining the effectiveness of an AI tool in terms of the quality of feedback were developed based on a systematic literature review and a survey of students and instructors. Given the small number of participants, the surveys were evaluated using qualitative empirical educational research methods. The results of this requirements analysis show that these technologies are considered to have great potential in supporting the individual learning process and that feedback quality is central in this.

This article describes the implementation of a teaching and learning scenario with the aim of integrating a Retrieval-Augmented Generation (RAG) chatbot into a laboratory-based course in fluid mechanics. In the Computational Fluid Dynamics (CFD) course, students learn how to use OpenFOAM, a software tool for solving complex fluid mechanics problems. Classes are held in a computer lab. Students work at individual workstations in front of a PC and the instructor demonstrates the relevant calculations and answers students' questions. In this setting, students are provided with a customized RAG chatbot as a “digital AI assistant.” The chatbot is trained to give students feedback that is comprehensible and conducive to learning. The results of this work will highlight factors for instructional design with regard to the effective use of LLM-based AI tools in instructional settings.

As part of the CrossLab project, LabCAR (Lab Constructive Alignment Recommender) was developed as an innovative tool that supports both students and teachers in the targeted design and implementation of learning processes. The system has a modular structure and enables the analysis and optimisation of learning outcomes (LOs), the coordination of learning and teaching activities, and the generation of well-founded recommendations for monitoring learning success. Students benefit from personalised suggestions for preparing for the defined learning outcomes and receive references to suitable learning resources, like digital laboratories. Teachers receive support in developing and adapting teaching methods and assessment strategies to promote the achievement of LOs. As part of the evaluation, various large language models (LLMs) were analysed in terms of their performance in the context of LabCAR. The study showed that all LLMs tested are fundamentally capable of fulfilling the defined objectives. However, significant differences in the quality and precision of the recommendations generated were found. The results underscore the importance of carefully selecting and adapting LLMs to meet the specific requirements of the education sector and ensure effective support for teaching and learning processes.

Cross-reality (XR) laboratories are becoming increasingly relevant in STEM education, enabling students to engage in experimentation beyond the re-strictions of traditional physical laboratory spaces. At the same time, large lan-guage models (LLMs) such as ChatGPT are rapidly entering higher education and raise new questions regarding competence development, learning practices, and educational integrity. The KICK 4.0 project addresses these developments by integrating AI-based natural language processing (NLP) tools into XR-based laboratory scenarios in fluid mechanics, with the aim of fostering both technical and reflexive competencies for a transforming world of work. This paper pre-sents a brief narrative literature review of the use of LLM/NLP tools in STEM learning contexts and analyses these findings through the lens of McLuhan’s Laws of Media (LOM; Enhance, Obsolesce, Retrieve, Reverse). This review suggests that LLMs may enhance creativity, reflection, and problem-solving; obsolesce highly scripted, instruction-driven laboratory formats; and retrieve di-alogical and inquiry-based learning practices. At the same time, risks of reversal become visible, including deskilling, bias, dependency, and ethical concerns. By combining the KICK 4.0 project with a LOM-guided literature analysis, this paper provides a structured reflection on how human–AI collaboration may re-shape laboratory education. The findings underline the need for critical AI liter-acy and balanced integration of LLMs in STEM learning environments.

Museums, as key institutions of non-formal learning, are experiencing rapid digital transformation shaped by emerging technologies. Among these, telepresence robots (TPRs) have attracted growing interest for their capacity to overcome physical and geographical barriers to participation. While immersive tools such as Virtual Reality (VR) and Augmented Reality (AR) are already widely employed to enhance engagement and visualization, the pedagogical and accessibility potential of TPRs remains underexplored.

This study investigates the opportunities and limitations of TPRs in museum education and assesses institutional readiness for adopting emerging technologies. Four research questions guided the work: (1) What is the level of readiness among museum educators and information professionals to integrate digital technologies? (2) How can technologies such as Artificial Intelligence (AI), VR, AR, and TPRs enhance learning processes? (3) What are the key challenges and opportunities for implementing TPRs in museum learning contexts? (4) What is the current technological capacity of museums to deploy robotic and immersive solutions for improved accessibility and inclusion?

A mixed-methods design was applied. First, a survey examined the readiness, experiences, and perceptions of museum educators concerning emerging technologies. Second, on-site observations and semi-structured interviews explored contextual factors affecting implementation practices, especially for TPRs. The survey, which included both closed and open-ended questions, covered digital competences, institutional support, infrastructure, perceived benefits and barriers, and ethical issues. Twenty-one professionals responded, and observations were conducted in twenty-two Estonian museums within the MUIS network.

Findings show high interest among educators in AI, VR, AR, and related tools. Most museums already employ digital solutions such as interactive screens, audio guides, and QR codes. While 81% of respondents expressed readiness to integrate TPRs, only 52% had prior awareness of their educational application. VR/AR tools were more familiar but often limited to temporary exhibitions. TPRs, by enabling remote real-time participation, were perceived as particularly inclusive for individuals with mobility or sensory impairments. Nevertheless, barriers included infrastructural constraints, insufficient funding, and limited staff competences. The study highlights the need for a sector-specific competence framework incorporating AI literacy, immersive design, and digital ethics.

By comparing TPRs with other emerging technologies, the study advances understanding of how museums can support educational objectives, foster inclusion, and drive digital innovation in non-formal learning.

Conclusions and recommendations emphasize that enhancing readiness for TPRs and related technologies requires infrastructural, organizational, and educational measures. Accessibility should be improved through ramps, elevators, barrier-free layouts, and optimized exhibition areas to facilitate robot navigation and visibility. Institutional capacity can be strengthened through designated charging zones, systematic staff training, and awareness-raising workshops. Additional practical skills, such as creating QR codes and audio guides, are needed. Pilot programs, teacher materials, and innovation funding should be developed to support sustainable and inclusive implementation of TPR-based and other technology-enhanced learning solutions.

This study contributes to the discourse on digital transformation in museums by clarifying the educational, accessibility, and organizational implications of telepresence robots and related technologies.

This paper presents the integration of Large Language Models (LLMs) and Physics-Informed Neural Networks (PINNs) into a Computational Fluid Dy-namics course based on deep learning and constructive alignment. The course included a lecture and a computer laboratory, conducted via Jupyter-Hub.NRW, enabling students to access remote Python environments and hardware. A survey of 14 students revealed that LLMs, primarily ChatGPT, were used mostly for programming assistance and debugging, while support for mathematical and physical understanding or generating examples was moderate, and planning activities was rare. Exam results indicate that per-formance differences were driven by cognitive levels as defined by the SOLO taxonomy rather than by AI versus traditional content, suggesting lim-ited intrinsic motivation toward AI. Students actively engaged with three AI Literacy Framework domains—Engaging with AI, Creating with AI, and Managing AI—while the fourth, Designing AI, was addressed through hands-on PINN tasks in lectures and laboratories. These activities exceed secondary education expectations of the AI Framework and represent a tertiary-level deepening of AI Literacy competencies. The findings highlight how AI can support numerics education, enhance learning outcomes, and provide a framework for developing advanced competencies in designing and critically evaluating AI systems in higher education.