(PDF) Enhancing Synchronous Collaborativ

PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The echoQuiz Approach Benedikt Brünner1[0009-0005-8484-9160] and Martin Ebner2[0000-0001-5789-5296] 1 Institute of Human-Centred Computing, Graz University of Technology, Graz, Austria 2 Educational Technology, Graz University of Technology, Graz, Austria

[email protected]

Abstract. This paper introduces echoQuiz, an open-source, AI-supported Audience Response System (ARS) designed for synchronous university (online) teaching with open-ended questions. The system follows a twophase interaction model: In the quiz phase, students/learners submit their responses and then rate their peers’ responses. In the echo phase, the instructor highlights one response for group reflection, with all responses remaining anonymous. To ease the interpretation of open responses, the lecturer can be assisted by an AI system during live sessions. Developed with an Educational Design Research (EDR) approach, echoQuiz was piloted in synchronous university courses with a total of 62 participants. Survey results show high motivation and moderate perceived learning gains. The findings suggest that free-text interaction, supported by AI, can enhance engagement and adaptability in digital classrooms. Keywords: Audience Response Systems, Artificial Intelligence in Education, Collaborative Learning, Motivation, Synchronous Teaching 1 Introduction 1.1 Background and Motivation In recent years, university-level synchronous online teaching has increasingly relied on video conferencing, especially during the COVID-19 pandemic, where synchronous teaching was widely adopted out of necessity [1]. While this shift has enabled accessibility and flexibility, it has also revealed serious challenges, including student active cooperation [4]. Past research has shown that interactivity is a key driver of student motivation and engagement in synchronous online settings. For example, in a recent workshop series on generative AI literacy, students highlighted interactivity as the most valued feature of the learning experience [2]. PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 2 Brünner, Ebner Existing Audience Response Systems (ARS) often rely on multiple-choice or single-choice formats [5], which may enable quick knowledge checks but fail to capture the learner’s thought process. In contrast, open-ended questions provide valuable insight into student understanding. However, they are more challenging to administer in synchronous settings because of the cognitive and practical demands of real-time processing, especially for instructors managing large groups. Despite the widespread availability of ARS tools, their integration of AI, particularly for real-time free-text processing, is still underdeveloped. Most systems are proprietary and focus on closed formats, limiting trust and flexibility. 1.2 Research Project Objectives and Questions The main goal of this project is to design, implement, and evaluate an AIsupported ARS that supports dynamic, open-ended interaction in synchronous (online) teaching. The research focuses on fostering learner motivation, perceived self-efficacy, and real-time adaptability. The system should support both students and instructors. It should increase students’ agency and interactivity and reduce instructors’ processing load through AI integration and transparent, open-source architecture. The following research questions guide this work: – How can an AI-supported ARS be technically and didactically implemented to support open, adaptive interaction in synchronous (online) teaching? – To what extent do students perceive this approach as engaging and helpful for their learning process? 2 Related Work 2.1 Audience Response Systems in Education Audience Response Systems (ARS) have evolved significantly over the past decades and are now a common feature in higher education for promoting interaction and engagement in large lecture settings. Their primary strength lies in increasing student motivation and perceived self-efficacy through active participation and higher attention during lectures [3]. Popular tools such as Mentimeter3 , Kahoot4 , Slido5 , and feedbackr6 offer user-friendly interfaces and immediate feedback loops, which can enrich the synchronous teaching experience. However, they have limitations that become obvious in advanced pedagogical contexts. Feedbackr only allows for limited interaction on other open-ended responses through +1 ratings and cannot focus on individual responses. Kahoot often enforces predefined answers and strict time limits, which can restrict deeper cognitive processing. Mentimeter and Slido offer a broad range of features, but are 3 https://www.mentimeter.com/, last accessed 2025-06-09 https://kahoot.com/, last accessed 2025-06-09 5 https://www.slido.com/, last accessed 2025-06-09 6 https://www.feedbackr.io/, last accessed 2025-06-09 4 PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Synchronous Collaborative Learning: The echoQuiz Approach 3 closed-source and pose challenges regarding displaying open-text responses effectively while balancing anonymity and participation tracking. Moreover, these platforms often lack transparency in how data is processed and stored, raising concerns around trust and GDPR compliance. 2.2 Challenges of Open-Ended Responses Open-ended responses are central to fostering deep learning and enabling students to articulate their understanding rather than merely selecting the most plausible option. This is especially important in university-level education, where topics are often complex and intended to encourage interpretation. From the instructor’s perspective, crafting good distractors for multiple-choice formats is time-consuming and may undermine the educational value of the question. Yet, integrating open-ended questions during synchronous sessions introduces several challenges. Students may be hesitant to speak or write when unsure, particularly if their contributions are publicly attributed. Anonymity becomes crucial for encouraging risk-taking and honest answers. At the same time, lecturers can face cognitive overload when reviewing dozens of diverse responses in real time, making it difficult to identify patterns or respond appropriately. 2.3 AI Integration in Digital Learning Tools Artificial Intelligence (AI) holds promise in addressing these challenges. AI can process large volumes of open-ended student answers quickly and could offer synthesized summaries, thematic clustering, or identification of common misunderstandings. However, the successful integration of AI into ARSs remains rare. Moreover, there is a lack of open-source ARS tools that transparently and pedagogically integrate AI. In educational settings, AI’s role should be that of an assistant, supporting instructors without replacing human judgment. Access to high-quality educational content is essential to ensure meaningful, contextaware AI feedback. Open Educational Resources (OER) play a critical role in this regard by providing accessible, adaptable materials that can be used to guide AI interpretation and processing coherently. 2.4 Collaborative and Active Learning Models The design of echoQuiz is inspired by established educational theories, particularly self-regulated learning [9] and constructivism [8], which emphasize learner agency, reflection, and active knowledge construction. In synchronous teaching, collaborative interaction fosters not just engagement but also metacognitive reflection, as students are encouraged to evaluate peer input and articulate reasoning. Furthermore, the concept of open learning environments (OLEs, [6]), which emphasizes learner autonomy, authentic problem-solving, and scaffolding to promote metacognitive processes in flexible, research-oriented environments, also guided the design. PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 4 Brünner, Ebner Real-time, open-text answers by students can foster rich dialogue and support learner agency. As Ebner notes, digital ARSs are especially powerful when they transform frontal lectures into interactive experiences that promote shared knowledge construction [3]. Such affordances are vital for higher education, where cognitive complexity and learner engagement demand tools that go beyond assessment to actively support the learning process. 3 Methodology 3.1 Research Approach: Educational Design Research (EDR) To guide the development of echoQuiz and explore its educational potential, we adopted the Educational Design Research (EDR, [7]) approach. EDR is particularly well-suited to technology-enhanced learning innovations because it explicitly aims to address complex real-world problems while generating theoretical insight. It does so through iterative cycles of design, implementation, and evaluation, in close collaboration between researchers and practitioners [7]. EDR differs from conventional research by simultaneously pursuing two goals: the development of practical solutions and the advancement of scholarly understanding. As described by McKenney and Reeves, EDR involves three core phases: (1) analysis and exploration, to understand the educational problem in context; (2) design and construction, to create and refine the intervention; and (3) evaluation and reflection, to study and improve both the intervention and the underlying theories [7]. This process is iterative, collaborative, and responsive to empirical findings and local needs. In our project, this methodology enabled the gradual development of both the technical features and the pedagogical model behind echoQuiz based on a specific didactic problem. The design cycles were informed by instructional design experts, student interaction data, and classroom observations. 3.2 EdTech System Design and Implementation Design Principles The development of echoQuiz was guided by three core educational principles: self-regulated learning, learner self-efficacy, and real-time adaptivity. These are foundational for learner-centered, collaborative knowledge construction, especially in synchronous online settings. The tool aims to empower students not only to respond but to co-direct lessons through open participation, while also reducing the instructor’s workload. Additionally, usability and accessibility were also central concerns. The platform adheres to clean HTML semantics (e.g., proper use of button elements), runs entirely in standard web browsers, and does not require students to install additional applications. This ensures accessibility across a broad range of devices and platforms. The functional scope was co-designed by a team of educational technology experts from TU Graz’s Future Learning, Analytics & AI for Teaching (FLAAIT) group. Feature prioritization was based on pedagogical value, ease of use, and gaps identified in the evaluation of existing ARS systems. PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Synchronous Collaborative Learning: The echoQuiz Approach 5 Technical Implementation Technologically, echoQuiz is built for ease of deployment and broad usability. The server is implemented in PHP7 and uses MariaDB8 for persistent storage. The frontend relies on HTML9 , CoffeeScript10 , and SASS11 for responsive design. Real-time interactions, including answer submission and updates, are powered by WebSocket technology using Nchan12 for nginx13 . The system architecture is illustrated in Figure 1. In terms of data protection, the system is GDPR-compliant. Instructors are unable to link individual responses to participants. A built-in reporting feature allows users to flag potentially harmful content. In such cases, a notification is sent to the instructor and a designated moderation team, who can then connect names and answers under strict policy controls. Instructors can immediately hide flagged content via a WebSocket call in the student’s frontend. fig1.jpg Fig. 1. System architecture of echoQuiz, showing lecturer and student interaction via browser, the role of nginx for routing, Nchan for real-time communication, PHP and MariaDB for backend logic and storage, and the integration of an external LLM for AI-assisted feedback. AI-Supported Open-End Answer Analysis The AI module processes openended text responses in real time to support instructors during live sessions. It uses a prompt-engineered call to an external large language model (currently via OpenAI’s gpt-4o mini API14 ) to classify and interpret responses. Each AI call includes the original question, the student’s response, and optionally additional instructional context. The AI serves exclusively as a support tool for lecturers. It provides a brief classification or contextualization of responses, which helps instructors identify 7 https://php.net/, last accessed 2025-06-09 https://mariadb.org/, last accessed 2025-06-09 9 https://html.spec.whatwg.org/, last accessed 2025-06-09 10 https://coffeescript.org/, last accessed 2025-06-09 11 https://sass-lang.com/, last accessed 2025-06-09 12 https://nchan.io/, last accessed 2025-06-09 13 https://nginx.org/, last accessed 2025-06-09 14 https://platform.openai.com/docs/guides/text/, last accessed 2025-06-09 8 PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 6 Brünner, Ebner patterns, misunderstandings, or opportunities for further discussion. The feedback is never directly transmitted to students, and no link between AI processing and student’s identity is possible for the AI. Interaction Flow and User Experience A typical echoQuiz session begins when the instructor opens a room and displays the QR code from /beamer via screen share. Students can then scan the code or navigate directly to echoquiz.eu. There, they enter the room ID and a name to participate and collect echoScore points for every interaction. The echoQuiz interaction is structured into two distinct phases: the quiz phase and the echo phase, see Figure 2. During the quiz phase, the lecturer posts an open-ended question, and students submit their responses anonymously before gaining access to and rating their peers’ answers. In the subsequent echo phase, the lecturer reviews the responses, assisted by AI-generated analysis and feedback hints if desired, and selects one to highlight for discussion. Students reflect on the highlighted answer. This two-phase model encourages spontaneous participation, peer learning, and immediate formative feedback within a single, integrated cycle. fig2.jpg Fig. 2. The two-phase structure of echoQuiz : During the quiz phase, students respond and rate answers; in the echo phase, the instructor highlights a selected response for reflection 3.3 Role of AI in echoQuiz While the student interaction in echoQuiz is centered around open-ended, peer-rated responses, lecturers are supported by a real-time AI feedback system that enables efficient moderation. This AI component is critical in connecting instant learner input with responsive teaching strategies. The AI system classifies and contextualizes student responses based on the prompt and question. It is not directly accessible by the learners and distracts their peer-focused experience, but assists the instructor by assisting at patterns, suggesting relevant topics, or highlighting potential misconceptions. This allows PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Synchronous Collaborative Learning: The echoQuiz Approach 7 the lecturer to identify noteworthy contributions more rapidly and choose appropriate answers for in-depth discussion. Figure 3 shows an example of how the system visually communicates AIsupported feedback to the instructor. While students only see peer content, the lecturer’s interface includes AI-evaluated classifications that help streamline decision-making in real time. fig3.jpg Fig. 3. Instructor view with AI-supported classification of open-ended student responses. Here, a student’s text response “Quality education” was correctly linked to SDG 4 from the AI assistants feedback. Instructors can choose to accept AI feedback or override it entirely. Importantly, all AI-generated content is kept private to the lecturer and is never sent to students by the platform. This ensures that the pedagogical judgment remains with the human teacher and that student trust and anonymity are preserved. The integration of AI in echoQuiz is not intended to replace teaching, but rather, it is designed to enhance instructional agility in real time. 4 Pilot Study 4.1 Participants and Setting The pilot implementation of echoQuiz took place in regular university courses at TU Graz. These included lectures in computer science education and maker education, conducted in both synchronous online and synchronous in-person formats. The total number of student participants was 62. Participation was anonymous and voluntary for enrolled students. However, students were motivated through echoScore points. These were awarded for engaging in the activity, answering questions, rating peer responses, and completing the post-session questionnaire. 4.2 Data Collection Data was collected using the in-app survey integrated directly into the echoQuiz platform. After each session, students were prompted to respond to five quantitative items on a 5-point Likert scale. No personal data or names were stored PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 8 Brünner, Ebner alongside their responses. While the platform supports qualitative response entry during learning activities, the evaluation in this phase focused solely on systematically collected quantitative feedback. 4.3 Survey Dimensions The survey focused on five key dimensions of student experience: 1. The quiz sparked my interest in the content 2. The interactive format was fun and motivating 3. I was able to effectively test my knowledge with the quiz 4. The feedback helped me understand my mistakes 5. I feel like I’ve learned something new from taking this quiz These items were custom-designed by the Teaching Academy at TU Graz as part of the university’s highest-level didactics training program. The goal was to capture both affective and cognitive dimensions of student experience in alignment with active learning and self-efficacy frameworks. 4.4 Data Analysis Methods The collected feedback data was analyzed using R15 . For each of the five questionnaire items, Figure 4 provides a visual representation of the response distributions. Fig. 4. Distribution of participant ratings per survey item, visualized as box plots with individual responses. 5 Findings 5.1 Student Engagement and Motivation The strongest positive response was observed for the item “The interactive format was fun and motivating”, which received a mean score of 4.07 (SD = 1.08, N = 59). This indicates that the majority of students perceived the echoQuiz activity as enjoyable and engaging. The high motivation rating aligns with prior findings that ARS tools can increase student attention and engagement during synchronous sessions. 15 https://www.r-project.org/, last accessed 2025-06-09 PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Synchronous Collaborative Learning: The echoQuiz Approach 5.2 9 Perceived Learning Effectiveness Students also indicated moderate agreement with statements related to perceived learning. The item “I feel like I’ve learned something new from taking this quiz” scored a mean of 3.85 (SD = 1.07, N = 52), while “I was able to effectively test my knowledge with the quiz” scored 3.42 (SD = 1.13, N = 53). These results suggest that students not only enjoyed participating but also experienced cognitive benefits, such as consolidating knowledge and recognizing learning gaps. 5.3 Helpfulness of Feedback The usefulness of feedback was evaluated through the item “The feedback helped me understand my mistakes”, which received a mean score of 3.40 (SD = 1.32, N = 52). Although this was the lowest of the five scores, it still indicates a generally positive perception. The relatively higher standard deviation suggests diverse experiences across sessions or individual feedback needs, which may reflect differences in how lecturers interpreted and used the AI-provided suggestions. 5.4 Interest in Content The statement “The quiz sparked my interest in the content” received a moderate mean of 3.52 (SD = 1.10, N = 62), suggesting that the interactive aspect of echoQuiz contributed to topic engagement, even among students who may not have initially been interested. This supports the system’s role in actively drawing students into the learning process. 6 Discussion 6.1 Interpretation of Findings The pilot results indicate that echoQuiz effectively contributes to students’ motivation and engagement. The highest-rated item, “The interactive format was fun and motivating”, with a mean of 4.07 out of 5, highlights the benefit of live interactivity in synchronous learning. This aligns with the pedagogical aim of fostering self-efficacy, which appears to be strongly supported through participatory and responsive teaching formats. One noteworthy finding was the relatively lower score for feedback (mean = 3.40). This can likely be attributed to the current design, in which AI-generated feedback is visible only to lecturers. As both students and instructors are still adapting to this new form of pedagogical support, the verbalization of feedback may have reduced its perceived impact. Future iterations could consider making selected AI-supported insights more transparent or interactive for students. It was somewhat surprising that motivation reached such a high score, especially given the minimal gamification and the experimental nature of the tool. This outcome suggests that the iterative, needs-driven development process guided by Educational Design Research (EDR) successfully identified and implemented features that resonate with learners. PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 10 6.2 Brünner, Ebner Implications for Teaching Practice echoQuiz offers several practical benefits for instructors working in synchronous settings. It reduces the cognitive load required to process student input by using AI to process open-ended responses and by allowing instructors to highlight specific answers during live discussions. These features foster spontaneous dialogue and help redirect teaching in real time based on students’ actual thinking. The system promotes a student-centered learning experience by encouraging learners to express their understanding rather than selecting from predefined options. In doing so, echoQuiz deepens engagement and addresses the risk of overreliance on generative AI by ensuring instruction is anchored in learner input. The option for anonymous participation promotes inclusivity by enabling less confident students to contribute without fear of judgment. In addition, the peer rating feature within echoQuiz offers an opportunity for fostering critical thinking and reflective learning, as students engage not only in expressing their understanding but also in evaluating and learning from their peers’ perspectives. 6.3 Strengths and Limitations of the Study A key strength of this study lies in its authentic integration into regular university courses, where echoQuiz was used under real-time teaching conditions. The in-app survey enabled seamless feedback collection, and the system sparked rich classroom discussions during synchronous online teaching, which were notably deeper and more student-driven than in previous semesters. This suggests that echoQuiz can effectively support active co-construction of knowledge, even in synchronous online contexts. One notable limitation of the platform is that the AI-supported feedback is available only for the instructor. A further limitation of this study is the relatively small sample size (N = 62). From a technical standpoint, the system functioned reliably, and students participated without hurdles. However, as the pilot was conducted at a technical university, it is possible that both students and instructors were more open to experimenting with educational technology. This may limit the generalizability of the findings and should be explored further in broader institutional contexts. 7 Conclusion and Future Work 7.1 Summary of Contributions This study introduces echoQuiz, a pedagogically sound and technically opensource Audience Response System (ARS) designed to facilitate open-ended and student-engaging interactions in synchronous teaching environments. Unlike most ARS tools, echoQuiz favors open-ended responses, provides AI-supported feedback for instructors, and offers a learner-centered teaching with peer feedback. Its core contribution lies in bridging the gap between structure and flexibility in digital education: it allows students to express their thinking freely, instructors PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 Synchronous Collaborative Learning: The echoQuiz Approach 11 to adapt in real time, and discussions to emerge even in large or online cohorts. echoQuiz supports at redirecting the focus back to students’ own understanding, reinforcing the value of human expression and dialogue in learning. 7.2 Reflections on the Role of AI Throughout the development and deployment of echoQuiz, AI support has proven helpful for teachers. It enables them to respond to complex, open-ended contributions from learners in real time. The integration was guided by a clear educational goal: to promote open, learner-centered interactions in real-time teaching contexts without overwhelming or replacing the teacher. In practice, lecturers reported that the AI-generated feedback helped them maintain high-quality discussions under time pressure and provided useful cues for selecting responses to discuss. The system demonstrates that AI can support spontaneous, responsive teaching rather than delivering pre-packaged content. From the lecturers’ perspective, this type of AI support was very valuable. Without it, implementing open questions in real time on this scale would not have been possible. Importantly, the AI never sent content to students without the lecturer checking its accuracy. Instead, it acted as a system that prepared the data for the teacher. This model offers a promising blueprint for integrating AI into (higher) education, not to automate teaching, but to complement it in a way that respects and enhances the human elements of the learning process. 7.3 Recommendations for Educators Educators interested in using echoQuiz can explore the open-source codebase at https://github.com/ed-tech-at/echoQuiz-eu or create their own sessions at https://echoQuiz.eu/admin. When designing questions, instructors are encouraged to avoid binary or overly closed prompts. Instead, questions should invite elaboration and reflection, which supports deeper learning and self-efficacy. While creating such questions may initially be unfamiliar, they open up richer opportunities for discussion and reveal a broader spectrum of student thinking. In synchronous settings, echoQuiz can be particularly effective when the instructor actively responds to answers, highlights responses on-screen, and uses AI-assisted feedback to spark dialogue. Providing contextual Open Educational Resources (OER) as background for the AI is important to further improve its support quality. 7.4 Directions for Future Research Future research should explore how echoQuiz functions across a broader range of universities and disciplines. Investigating long-term adoption and comparing different pedagogical uses could yield further insights into its effectiveness and adaptability. PREPRINT Finally published in: Brünner, B., Ebner, M. (2026). Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The EchoQuiz Approach. In: Auer, M.E., Toth, P. (eds) Innovation via Collaborative Learning in Engineering Education. ICL 2025. Lecture Notes in Networks and Systems, vol 1847. Springer, Cham. https://doi.org/10.1007/978-3-032-18885-4 2 12 Brünner, Ebner On the technical side, upcoming developments include the integration of open-weights large language models and additional LLM providers. Planned features also include answer clustering for large-scale classes and AI-generated question suggestions based on context. These enhancements aim to further support instructors while maintaining transparency and pedagogical intent. Ultimately, this work contributes to a growing body of research on how AI can enrich education, not by automating instruction, but by empowering teachers and students to engage more meaningfully with each other and with the subject matter. Acknowledgement. This research was done as part of the ”FutureDEAL Future of Digital Education and Learning” initiative within the doctoral program ”Bildungsinnovation braucht Bildungsforschung”, which is supported and partially funded by the Austrian Federal Ministry of Education and Austrian Federal Ministry of Women, Science, and Research. References 1. Brünner, B., Findenig, K., Ebner, M.: Digital learning during covid-19. a systematic review and meta-analysis of distance learning at universities in austria. In: Bastiaens, T. (ed.) Proceedings of EdMedia + Innovate Learning. pp. 980–988. AACE (2024), https://www.learntechlib.org/primary/p/224617/ 2. Brünner, B., Schön, S., Ebner, M.: From gretel to strudelcity: Empowering teachers regarding generative ai for enhanced ai literacy with collectivegpt. Education Sciences 15(2), 206 (2025). https://doi.org/10.3390/educsci15020206 3. Ebner, M.: Audience-response-systeme. FNMA Magazin (04/2015), 10–11 (2015), https://fnma.at/content/download/758/2744 4. Ferri, F., Grifoni, P., Guzzo, T.: Online learning and emergency remote teaching: Opportunities and challenges in emergency situations. Societies 10(4), 86 (Nov 2020). https://doi.org/10.3390/soc10040086 5. Haintz, C., Pichler, K., Ebner, M.: Developing a web-based question-driven audience response system supporting byod. Journal of Universal Computer Science 20(1), 39– 56 (2014). https://doi.org/10.3217/jucs-020-01-0039 6. Hannafin, M.J., Land, S.M., Oliver, K.M.: Open learning environments: Foundations, methods, and models. In: Reigeluth, C.M. (ed.) Instructional-design theories and models: A new paradigm of instructional theory, vol. 2, pp. 115–140. Lawrence Erlbaum Associates, Mahwah, NJ (1999) 7. McKenney, S., Reeves, T.C.: Educational design research for relevant & robust scholarship. Journal of Computing in Higher Education 37(2), 614–638 (2025). https://doi.org/10.1007/s12528-025-09456-2 8. Papert, S.: Situating constructionism. In: Harel, I., Papert, S. (eds.) Constructionism: Research Reports and Essays, 1985–1990, pp. 1–12. Ablex Publishing Company, Norwood, NJ (1991) 9. Zimmerman, B.J.: Investigating self-regulation and motivation: Historical background, methodological developments, and future prospects. American Educational Research Journal 45, 166–183 (2008). https://doi.org/10.3102/0002831207312909

(PDF) Enhancing Synchronous Collaborative Learning with AI-Supported Audience Response Systems: The echoQuiz Approach