1ο Διεθνές Επιστημονικό Συνέδριο. Artificial Intelligence and Innovation in Education: Ethical and Technological Dimensions (Thessaloniki: 12-13/12/2025 )

Μόνιμο URI για αυτήν τη συλλογήhttps://repository.ihu.gr/handle/11544/48409

Περιηγούμαι

Τώρα δείχνει 1 - 20 από 89

AI and AE: Aristotelian Virtue Ethics at the Core of the Ethical Use of Artificial Intelligence in Lower Secondary Education
(2025-12-12) Bakali, Olympia; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The rapid integration of Artificial Intelligence (AI) into education constitutes not merely a technological development but a profound anthropological and pedagogical transformation reshaping learners’ relationships with knowledge, truth, judgment, agency and community. Although contemporary institutional frameworks (AI Act, EU Guidelines, National Documents) emphasize risk mitigation, human oversight, fundamental rights, they largely remaining external to ethical formation in lived school practice. This study addresses a clear scientific gap: the absence of a coherent pedagogical model capable of foregrounding how AI ethics is understood and explicitly enacted in classrooms. It is, thus, situated within the genre of conceptual research as it advances theory through normative reasoning, philosophical analysis and interpretative engagement with institutional and educational discourse rather than empirical investigation. The main aim of the study is to propose a reflective, virtue-based pedagogical and philosophical model for AI Ethics Education in secondary education that integrates critical understanding of AI into the core of the educational process addressing AI ethical dilemmas and supporting the holistic development of the school community. It seeks to: a. synthesize Aristotelian virtue ethics with contemporary AI ethics discourse and pedagogical theory, b. propose the in situ co-creation of an Internal School Code of Digital Ethics (ISCDE) as a humanistic guide for cultivating Digital Phronēsis-Prudence, c. elucidate the Code’s pedagogical logic as a tool for responsible AI use, teacher professional empowerment and collective ethical culture formation. The model integrates Aristotelian virtue ethics -mesotēs, hexis proairetikē and phronēsis- (virtue as mean, habituation by deliberative choice, practical prudence) drawing on Aristotle’s Nicomachean Ethics (NE). This foundation is extended through Techno-moral virtues theory, Critical Pedagogy, Multiple Intelligences, Social Constructivism, Holistic Learning and the methodological use of Critical Discourse Analysis and Thematic Analysis. Two school textbooks -the Anthology of Philosophical Texts and Euripides’ Helen serve as pedagogical anchors. The above Aristotelian theory alongside the ontological dilemma of εἶναι–φαίνεσθαι (being-appearing) in Helen is reinterpreted in relation to AI verisimilitude, demonstrating the central role of humanities subjects in cultivating ethical judgment within digital education. Methodologically, the model follows a three-phase conceptual process: Critical Discourse Analysis of official AI texts by teachers as institutional moral literacy; thematic synthesis of values re-signified as virtues; classroom “scenarios” enabling experiential AI habituation. Institutional values are transformed into digital virtues articulated as means between AI deficiency and excess and embodied in the ISCDE, a co-created ongoing pedagogical artifact. An example is given on Digital Liberality, where prompt sharing and attribution practices shift AI use toward communal epistemic contribution. Expected outcomes are multilevel: students cultivate Digital Prudence, moral judgment, balanced AI use and critical agency; teachers develop institutional literacy and professional confidence; schools foster a shared ethical culture of AI governance. Policy recommendations to the Institute of Educational Policy involves advocating the institutionalization of AI Ethics Education as a cross-curricular priority, the integration of Aristotelian ethics and humanities-based reflection into curricula and the strengthening of teacher institutional literacy. The originality of the study lies in defining Digital Phronēsis as a new educational competence for human-centered education in the AI era.
Earthquaker-AI: A Retrieval-Augmented Generation Framework with Rubric-Based Assessment for Primary School Earthquake Education
(2025-12-12) Kokkinou, Xanthi; Mizeli, Chaido; Koulaxidou, Nafsika; Delianidi, Marina; Diamantaras, Konstantinos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—This paper presents Earthquaker-AI, a hybrid educational framework that builds upon a previously implemented educational robotics project by integrating a conversational artificial intelligence assistant based on Retrieval-Augmented Generation (RAG), aiming to enhance earthquake preparedness and conscious action among primary-school students. The system extends the earlier award-winning STEM project Earthquaker, which received 3rd place in the Hellenic WRO Competition (2020), and moves from mechanical simulation through Lego WeDo2 activities to cognitive and metacognitive processing. The robotics component, developed in the original Earthquaker project, employed Lego WeDo2-based automation to simulate seismic response, allowing students to interact with sensors and actuators that function as tangible representations of core protective actions. At the cognitive level, the conversational assistant adopts a Retrieval-Augmented Generation architecture, in which evidence retrieval precedes language generation, so that responses are grounded in official educational material and aligned with the intended pedagogical framework. The assistant serves a dual role: it operates as a guided learning mechanism that aligns students’ responses with institutional safety guidelines, while simultaneously providing rubric-based verbal feedback that supports self-regulated learning and the development of calmness under emergency conditions. Earthquaker-AI follows a progressive learning trajectory aligned with students’ cognitive development across age stages. In the early grades, the emphasis is placed on basic recognition of earthquake-safety actions through simple multiple-choice questions, in which students select the most appropriate answer, supporting orientation toward the recommended behavior during seismic events. Assessment at this stage is conducted using a two-dimensional rubric focused on action recognition and emotional regulation. In middle grades, learning activities become more demanding, requiring students to identify the correct sequence of actions through multiple-choice questions, evaluated using a three-axis rubric that captures organized thinking and decision-making. In upper grades, the approach shifts from recognition to verbal production, with students providing short written responses assessed through a four-dimensional rubric. The additional dimension of clarity of expression reflects increasing metacognitive maturity, as it requires structured reasoning, justification, and precise articulation. The system also includes a dedicated dialogic module that leverages RAG to retrieve and synthesize evidence-based answers. Student queries are semantically matched with selected excerpts from official earthquake-safety guidelines, from which pedagogically safe and accurate responses are generated. Experimental evaluation demonstrates high answer groundedness (0.84) and accuracy (0.85), together with a low hallucination rate (0.07), indicating stable and evidence-aligned behavior under the evaluated experimental conditions. Overall, Earthquaker-AI presents an integrated educational approach to earthquake preparedness, bringing together hands-on engagement, information processing, and reflective practice. The sequence of observation, physical interaction, interpretation, and verbal articulation provides a holistic learning experience. At the same time, the combined use of robotic processes, analytic rubrics, and artificial intelligence promotes technological literacy, self-regulation, and responsible use of digital systems in primary education, contributing meaningfully to earthquake preparedness and the early development of crisis-management skills.
The training needs of adult educators in the learning process using artificial intelligence
(2025-12-12) Zygouris, Fotios; Vlachou, Vasiliki; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract - The rapid development of artificial intelligence ( AI ) is redefining both educational environments and the demands placed on adult educators, making their training critical for the effective use of new information and communication technologies in the learning process. Adult education, as a field with specific characteristics—such as diversity of experiences, self-directed learning, and orientation to labor market needs—requires educators who are prepared to integrate AI in a pedagogically valid and ethically responsible way ( Knowles , 1984 ; Merriam & Bierema , 2014 ). However, research shows that most adult educators feel inadequately prepared to use AI tools , often due to insufficient digital pedagogical training and limited knowledge regarding the possibilities and risks of new applications ( UNESCO , 2023 ). The purpose of this paper is to highlight the need for systematic and targeted training of adult educators for the use of AI in the learning process. Since AI has the potential to support the personalization of learning, the provision of immediate feedback and the formation of adaptive learning paths, its integration facilitates the enhancement of the participation and autonomy of adult learners ( Holmes and al ., 2021; Popenici & Kerr , 2017). However, the technological and pedagogical competence of educators is a determining factor for the success of any digital intervention and digital competencies of teachers ( Mishra & Koehler , 2006; Redecker , 2017). The analysis of the literature and the findings of recent studies indicate that professional development programs should include: (a) understanding basic principles of AI , (b) development of digital pedagogy, (c) practical application of AI tools in real educational environments, and (d) training on issues of ethics, transparency, data protection and algorithmic bias ( Akgün & Greenhow , 2022; Long & Magerko , 2020). At the same time, the adoption of collaborative learning models among educators, formative assessment of skills and continuous renewal of educational materials are proposed, as AI is evolving at a very rapid pace. The conclusions of the paper highlight that the training of adult educators can no longer be considered a complementary activity, but a basic prerequisite for the quality and effectiveness of education in conditions of digital transformation. Cultivating AI competencies in adult educators not only enhances their pedagogical competence, but also ensures the creation of inclusive, innovative and reflective learning environments, suitable for the needs of adult learners in the 21st century. Finally, directions are proposed for future research that will examine the benefits, risks and best practices of integrating AI into adult education, with the aim of developing efficient training models.
Case Adaptive e-Learning Systems for Sustainable Healthcare: Personalized Green Lean Six Sigma Training for Hospital Staff
(2025-12-12) Vasileiou, Anastasia; Sfakianaki, Eleni; Tsekouropoulos, Georgios; Hoxha, Greta; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract —This paper proposes a conceptual framework for designing adaptive e-learning systems that support hospital personnel in developing competencies in Green Lean Six Sigma (GLSS). GLSS is an integrated methodology that combines Lean process optimization, Six Sigma quality improvement, and environmentally sustainable operational practices. As hospitals confront the growing need to enhance clinical efficiency while reducing their ecological footprint, they require training approaches capable of equipping diverse professional groups with the skills needed to address these dual priorities. The aim of this study is to demonstrate how artificial intelligence (AI)–enabled educational technologies can personalize GLSS learning pathways, strengthen learner engagement, and increase the transfer of knowledge and skills to everyday clinical contexts. The framework is structured around three main components. The first component involves detailed learner profiling and an initial diagnostic assessment to establish baseline knowledge, identify role-specific requirements, and capture individual learning preferences. The second component uses real-time learning analytics to dynamically adapt instructional materials by adjusting difficulty levels, modifying activity sequences, and calibrating the level of scaffolding in response to learner performance and progression. The third component integrates AI-driven assistants built on retrieval-augmented generation (RAG), which deliver tailored feedback, case-based explanations, error analyses, and contextually grounded examples drawn from actual hospital workflows. Through these mechanisms, the system aims to promote core GLSS capabilities, including waste recognition, root-cause analysis, process redesign, and environmentally responsible decision-making. The instructional design includes micro-learning modules, interactive clinical scenarios, simulation-based exercises, and applied activities using anonymized hospital datasets. These elements are intended to help learners connect GLSS principles with everyday responsibilities, such as reducing unnecessary consumption of materials, improving patient flow, and supporting sustainable resource use. Ethical considerations are embedded throughout the system’s architecture, particularly in relation to data transparency, responsible AI use, and strategies to minimize algorithmic bias in adaptive personalization. Expected outcomes of the proposed approach include increased learner motivation, measurable gains in GLSS-related knowledge, and observable improvements in sustainability practices across clinical departments. By tailoring content to learners’ roles and progress, the system has the potential to shorten training time while enhancing relevance and practical applicability for various healthcare professionals. The evaluation plan incorporates pre- and post-training assessments, detailed analytics on learner interactions, and operational metrics such as waste reduction indicators and measures of environmental impact. Overall, this abstract outlines how adaptive e-learning technologies-when integrated with GLSS principles-can support the transition toward more sustainable healthcare systems. The combination of adaptive instructional design, AI-based personalization, and data-driven feedback represents a coherent model for preparing healthcare staff to address environmental challenges while improving operational performance. Future research will focus on developing a working prototype, assessing its educational effectiveness, and examining its influence on sustainability outcomes in real clinical environments.
Evaluating Gamma App as an AI-Powered Authoring Tool for Educational Content Creation
(2025-12-12) Vakalis, Alkis; Dragogiannis, Konstantinos; Papantoni, Iliana; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract —This study investigates teachers’ acceptance and post-training use of Gamma App—an AI-powered platform for educational content creation—among secondary education teachers who attended a related professional development seminar. A total of 18 teachers from various subject areas participated, evaluating both the training and the tool’s educational value, while key constructs grounded in the Technology Acceptance Model (TAM) were also measured. Results indicate that most participants perceived the seminar as clear and satisfactory, although some rated the duration and/or post-training support as only moderate. Following the seminar, approximately 38.9% of participants (7 out of 18) reported using Gamma App in their teaching practice, primarily for creating lesson presentations, on a weekly or monthly basis. These users reported high perceived usefulness and perceived ease of use (mean scores 4.33/5 and 4.55/5, respectively), along with positive satisfaction and a strong intention to continue using the tool. Internal consistency for the perceived usefulness and perceived ease-of-use scales was high (Cronbach’s α > .90), indicating reliable measurement. In contrast, 11 teachers did not use Gamma App after the seminar. The most prominent barrier was lack of time, followed by technical constraints and infrastructure limitations. Perceived lack of support, ethical concerns, and perceived relevance to the subject area were rated as less influential barriers. Moreover, most participants expressed interest in further professional development on AI tools, suggesting a willingness to pursue ongoing growth in digital competencies. Overall, Gamma App was evaluated positively by teachers who actively tried it; however, time availability, sustained support, and enabling conditions appear necessary to facilitate broader adoption in everyday educational practice.
Beyond the Chatbot: Co-Learning and Co-Teaching through a Dual-Persona Generative-AI Assistant
(2025-12-12) Mizeli, Chaido; Delianidi, Marina; Diamantaras, Konstantinos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—In this paper we present a generative-AI application developed to support both teachers and students in secondary education. The system employs two Large Language Models (LLMs) – Gemini and DeepSeek, and a Small Language model (SLM) – Gemma, integrated within a Retrieval-Augmented Generation (RAG) framework to create a pedagogically grounded, Greek-language assistant capable of adapting its reasoning and communication style to the user’s role. Unlike conventional chatbots, the assistant introduces Pedagogical Persona Switching, a dual-role mechanism that enables the same AI model to act as both a teaching companion and a learning guide. When interacting with a student, the assistant produces accessible, curriculum-grounded explanations that promote conceptual understanding while remaining faithful to the officially approved educational sources. When addressing a teacher, the assistant generates diverse artifacts, including assessment items, classroom activities, and structured lesson plans. This dual-persona approach bridges knowledge retrieval with pedagogical creativity, enhancing differentiated learning. Utilizing a RAG paradigm tailored to the Greek educational domain, the architecture segments official textbooks into coherent units. Enriched with specific metadata, these units preserve curricular structure and instructional context, demonstrating how generative AI optimizes modern instructional design. These units are represented by domain-specific sentence embeddings fine-tuned on Greek semantic similarity tasks. This linguistic adaptation ensures precise meaning alignment between user queries and content segments, maintaining fidelity to the official curriculum. All representations are organized within a retrieval index, guaranteeing transparent and source-grounded generation. The initial case study focuses on Home Economics in Greek lower-secondary education (Grades A–B), a cross-disciplinary subject that integrates elements of economics, health education, and social responsibility. The assistant has been developed to support both learners and educators in complementary ways. In future classroom implementations, students will be able to use it to clarify key concepts such as financial literacy, resource management, and healthy living, while teachers could employ it to design authentic instructional materials, formative assessments, and classroom activities aligned with the official curriculum. Home Economics (HE) was chosen as the pilot domain because it is a non-STEM, value-oriented subject focused on early adolescents, emphasizing ethical reasoning, social awareness, and everyday decision-making. This unique context allows us to explore generative AI's pedagogical potential beyond traditional core subjects, particularly its capacity to foster reflective reasoning and civic responsibility in young learners. The project introduces an innovative generative framework combining three key elements: (a) a localized, curriculum-specific Retrieval-Augmented Generation (RAG) environment, (b) semantic educational text segmentation, and (c) adaptive persona-driven prompting. More than just a technical solution, this framework demonstrates how generative AI can be pedagogically and linguistically aligned with national curricula and local educational contexts. The resulting model offers a reproducible foundation for future AI integration into authentic learning settings, supporting teacher agency, learner engagement, and values-based secondary education. The study elevates the concept beyond a simple chatbot, proposing a structured, contextually adaptive framework for pedagogical generative assistants that effectively bridge technology, curriculum, and human learning.
AI into Foreign Language Learning: Pathways to Innovation and the Challenge of Integrity
(2025-12-12) Kotsoni, Elissavet; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—Artificial intelligence (AI) is reshaping the landscape of foreign language education, offering new pathways for learner engagement, instructional innovation, and teacher development. This presentation synthesizes recent empirical and review studies on the integration of generative AI tools, particularly ChatGPT, into foreign language learning and teacher education. Findings demonstrate that AI-mediated instruction enhances language achievement, supports self-regulated learning, and fosters learner motivation through personalized feedback and authentic conversational practice. Systematic reviews further reveal that AI is increasingly embedded in teacher training programs, prompting both pedagogical opportunities and critical discussions around ethics, assessment, and professional preparedness. In addition, teaching on computer-assisted language learning (CALL) highlights AI’s role in customizing instruction and expanding access to language resources. On the contrary, the systematic unwise exploitation of AI raising questions when it comes to ethics and the creation of a lazy generation. Recapitulating, these findings reveal that AI has significant promise in store for advancing foreign language learning and teaching, though challenges remain in ensuring equitable access, maintaining academic integrity, and preparing educators to critically integrate AI into their syllabus. This synthesis contributes to emerging debates about the future of language education in an era of smart technology.
Leveraging an AI-Based Digital Assistant: A Pilot Implementation in Early Childhood Communities of Learning and Practice within the Directorate of Primary Education of the Third District of Athens
(2025-12-12) Foti, Paraskevi; Σχολή Μηχανικών, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract —This paper presents the results of a pilot intervention implemented in early childhood education of the Directorate of Primary Education of Athens during the academic years 2023–2024 and 2024–2025. The aim of the study was to investigate how the use of digital assistants can support inclusive and personalized learning in early childhood education. The conception of the intervention was also informed by the researcher’s experience through participation in an official European Commission initiative focusing on the systematic integration of innovative digital tools into educational practice. This experience contributed to the design of an intervention that combined pedagogical principles, digital practices, and the everyday needs of kindergarten teachers. A mixed-methods research approach was employed to evaluate the intervention, incorporating both quantitative and qualitative data. In the initial phase, online informational and training sessions were conducted in order to familiarize kindergarten teachers with the functionalities of the digital tools and their pedagogical potential. Subsequently, the tools were organically integrated into the daily classroom routine, mainly in activities related to the organization of the school day, transitions between activities, and the enhancement of participation for all children, particularly those requiring additional support to meet the demands of the school environment. The implementation was accompanied by continuous feedback, addressing both practical issues and teachers’ reflective engagement with the use of digital assistants. Quantitative data were collected through a questionnaire administered after the completion of the intervention, focusing on perceived usefulness, ease of use, pedagogical value, and the extent to which the tools supported classroom organization. Qualitative data were derived from teachers’ reflective records and open-ended questionnaire responses. The analysis of the data provided insight into how teachers experienced the process and the changes they observed in children’s participation and attitudes. The findings indicated that kindergarten teachers responded positively to the implementation of digital assistants, perceiving them as significantly contributing to more effective classroom organization and the management of daily routines. At the same time, teachers reported that children responded with enthusiasm, while the visual support and clear structuring of activities were particularly beneficial for students who require stable and predictable learning environments. Overall, the study highlights that the use of digital assistants in early childhood education can function as a supportive mechanism for promoting inclusion and personalized learning, especially when accompanied by appropriate professional development and systematic support. The findings are aligned with the broader directions of the European Commission regarding the digital transformation of education and underscore the need for further research with larger samples and across diverse educational contexts. A concise and factual abstract is required (maximum length of 500 words). The abstract should state briefly the purpose of the research, the principal results and major conclusions. References should be avoided, but if essential, then cite the author(s) and year(s). Also, non-standard or uncommon abbreviations should be avoided, but if essential they must be defined at their first mention in the abstract itself.
The Educator as an AI-based Mediator in the Era of Education 5.0: Reducing Educational Inequalities
(2025-12-12) Kota, Eris; Nikolidakis, Symeon; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The advent of new technologies and their integration into educational settings have carried a great promise to transform teaching and learning, to foster creativity, to advance critical thinking, and to democratize knowledge. The introduction of Artificial Intelligence (AI) applications in education is seen by some as a blessing and by others as a curse. The concerns are intense, mainly due to the fact that minors are involved, whose personalities are shaped within the educational process. This does not mean that we should ban the introduction of AI tools in education, but that we must limit their application with the basic aim of improving the educational process without losing sight of its fundamental goal, which is to shape responsible and conscious citizens and cultivate ethical thinking. This requires the educator's role to become that of a coordinator and advisor and not that of the transmitter of knowledge, moving toward an era of future education in which teachers and AI technologies work together. Keywords: Artificial Intelligence, Educational Inequality, Educator Mediation
From Algorithm to Human: Pedagogical Autonomy, Critical Literacy, and Bridging Social Inequalities in the Age of Artificial Intelligence
(2025-12-12) Nikolidakis, Symeon; Tromara, Sophia; Pliogou, Vassiliki; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract - Artificial Intelligence (AI) has become an integral part of social life, including the field of education, where in most cases it is used as a tool for personalisation, efficiency, and innovation, opening thus new pedagogical pathways. However, AI implementation is accompanied by significant legal, ethical, and pedagogical concerns, raised by core features that characterize AI systems, such as the intensive processing of personal data (including sensitive data), and the subsequent uncertainty in terms of the reliability, fairness, and accountability of their outcomes. Those issues become more critical within educational settings, as minors are involved and as education assumes a crucial role in shaping agency, autonomy, critical thinking, and democratic citizenship. In line with these, the present theoretical paper examines the application of AI in education, by applying the perspective of critical literacy, pedagogical autonomy, and social inequality, while taking into account the EU regulatory framework (AI Act), which imposes rules of transparency and respect for human dignity, aiming to prevent manipulation and social exclusion. Finally, the aim of this theoretical paper is to contribute to contemporary debates that highlight that the educational value of AI depends highly on the conditions of its use, and suggest that AI should only be used as a human-centred and ethically-grounded supportive tool that has the power to enhance teachers’ role and deliver meaningful human interaction, with the ultimate goal of shaping responsible citizens with well-developed critical thinking skills.
Studies Designed as a Game: Rethinking Curriculum through Play, Flow, and AI co-agency
(2025-12-12) Katsenou, R.; Garneli, V.; Mazi, I. A.; Goudelis, L.; Christodoulopoulou, E.; Koustas, M.; Poulimenou, S-M.; Deliyiannis, Ι.; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract —This conceptual paper proposes a framework for transforming higher education distance learning by reconceptualizing an entire Master’s curriculum as a Serious Video Game. Rather than gamifying isolated modules, the proposal restructures the full academic program as a coherent game world. AI-powered tools are integrated to lower the technical threshold of game development, enabling novices to co-create complex content through assisted programming and asset generation. Learning unfolds within a narrative-driven environment structured around role-based collaboration, player development, and a dynamic progression system. An inverted scoring model—beginning at its maximum and gradually decreasing—introduces productive tension that mirrors gameplay logics while sustaining engagement. A central narrative of a fragmented “book of knowledge” positions students as active agents who restore disciplinary coherence by confronting domain challenges. Player agency, including character customization and thematic choice through in-game currency, supports identity formation and professional orientation. The framework synthesizes Huizinga’s “Magic Circle” and Csikszentmihalyi’s “Flow” with a posthumanist understanding of Human–AI co-agency, positioning AI not as a tool of assistance but as a co-constitutive actor in learning. In doing so, the model contributes to Education 5.0 by operationalizing symbiotic Human–AI collaboration within an accredited academic structure. Methodologically, the study employs a research-through-design approach grounded in long-term pedagogical praxis to articulate a scalable model that preserves full academic equivalence through formal LMS certification of in-game progression.
Investigation of the Use of Artificial Intelligence Tools by Primary School Teachers
(2025-12-12) Diamantopoulos, Ilias; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The present study aims to investigate teachers’ familiarity with artificial intelligence, as well as their perceptions and attitudes toward its integration into the educational process. Specifically, the study examines teachers’ level of knowledge and skills related to the basic principles and applications of AI, the frequency of AI tool use in teaching practice compared to everyday life, the types of AI tools most commonly used, and potential differences based on gender, years of teaching experience, and training background. In addition, the study explores teachers’ experiences with AI use both in their personal lives and in instructional settings. To achieve these aims, a quantitative research design was employed using a self-report questionnaire that included mainly closed-ended questions, along with a limited number of open-ended questions that were analyzed qualitatively. The sample consisted of 190 primary school teachers, kindergarten teachers, and teachers of various specialties. Convenience sampling was used, as participants were selected based on their availability and accessibility. The findings indicate that the majority of teachers primarily use general-purpose AI tools, such as ChatGPT, Gemini, and Canva. Although these tools are not specifically designed for educational purposes, teachers adapt them to support instructional tasks. In contrast, specialized AI tools explicitly developed for educational use—such as Magic School, Curipod, Diffit, and Teachy—were mentioned rarely or not at all, suggesting that the pedagogical potential of AI has not yet been systematically exploited. Overall, the results highlight a clear need for targeted teacher training focused on the pedagogical and critical use of AI, increased familiarity with specialized educational AI tools, and the cultivation of ethical and responsible attitudes toward AI integration. The findings are expected to contribute to the ongoing scientific dialogue on the educational use of artificial intelligence and to emphasize the importance of well-designed professional development initiatives that support the creative and effective integration of AI into the teaching process.
Evaluating Large Language Models on Greek Primary School Mathematics Problems: Accuracy, Reasoning, and Educational Implications
(2025-12-12) Volakakis, Argyris; Tzivinikou , Sotiria; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract — The rapid advancement of Large Language Models (LLMs) such as the GPT family is reshaping expectations for automated tutoring, formative feedback, and assessment in education. Despite impressive progress in natural language understanding, the ability of these models to perform reliable mathematical reasoning, especially in non-English languages and primary-level educational contexts, remains insufficiently examined. This study investigates how accurately and consistently current LLMs solve mathematical problems from the Greek primary school curriculum and how their reasoning patterns relate to task difficulty. A dataset of 70 mathematics problems was created based on the 3rd- and 5th-grade Greek mathematics textbooks. The problems are categorized by grade, topic (e.g, number sequences, fractions, geometry), and a difficulty scale (1–4), which is validated by three experienced educators and two LLMs. Using a Python-based evaluation pipeline, each problem was submitted in the Greek language to four widely known OpenAI models, i.e., GPT-3.5-turbo, GPT-4o, and GPT-4.1-mini, and GPT-5-mini, under two prompting conditions: (i) the LLM to provide a direct answer and (ii) the LLM to provide the complete step-by-step reasoning. All answers were forwarded to a dedicated Python module to assess accuracy based on the corresponding ground truth in the database, as well as time and cost spent per model. Additionally, step-by-step responses were analyzed using an LLM-as-a-Judge approach, where the GPT-5 model rated the reasoning quality, completeness, intermediate calculations, understanding, and consistency of each model, besides accuracy. This dual analysis offers an initial exploration of hybrid methods for assessing AI reasoning in educational tasks. The evaluation revealed systematic performance differences across grade levels, mathematical domains, and difficulty levels. The overall accuracy is higher on 3rd-grade problems and computationally oriented topics, reaching approximately 90%, while it decreased substantially for problems with greater difficulty or more complex reasoning. As task difficulty increased, accuracy declined by about 8% for GPT-5-mini and by more than 40% for GPT-3.5-turbo, indicating a sensitivity to cognitive demand, particularly for earlier-generation models. The contribution of this work lies in showing where and why these differences in LLM performance occur and how they relate to specific task types. Beyond the quantitative findings, the study considers key pedagogical and ethical issues associated with deploying LLMs in primary mathematics education. This research proposes a small-scale yet reproducible framework for evaluating LLMs’ mathematical reasoning in non-English elementary settings. Future work will expand the dataset, examine additional LLMs beyond OpenAI’s models, and evaluate the quality of their tutoring capabilities in mathematics. These analyses provide initial evidence relevant to the design of LLM-based tutoring tools for elementary mathematics. The findings aim to inform both developers designing educational AI systems and educators seeking to integrate such technologies responsibly into classroom practice.
From Text to Image: Introducing Artificial Intelligence in the Teaching of Literature in Vocational High School
(2025-12-12) Tsiklias, Foivos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract — This paper investigates the use of text‑to‑image Artificial Intelligence tools in teaching Modern Greek literature within Vocational High Schools (EPAL) in Greece. Drawing on contemporary theories of multimodal literacy, human‑centered AI, and the specific pedagogical conditions of vocational schooling, it examines how AI‑supported visualization may enrich students’ engagement with literary texts and support deeper awareness of the literary technique of description. The educational intervention unfolded across three interconnected phases. First, students engage in close reading and systematic analysis of descriptive strategies in K. Ouranis’ text “The old Athenian carnival,” identifying key features and producing initial visual interpretations. The second phase introduces text‑to‑image AI applications, through which students transform selected passages into prompts and generate corresponding images. Collaborative prompt construction and comparison between textual and visual outputs encourages interpretive precision and introduces students to basic principles of AI‑mediated representation. The final phase emphasizes on critical reflection and production of written speech, as students present their prompts and images, discuss the interpretive dynamics between text and visualization, and compose reflective pieces on their learning process. Data sources include student artifacts, classroom observations, teacher field notes, and group presentations. A qualitative thematic analysis suggests that the intervention supported more attentive engagement with literary description, fostered emerging forms of digital and AI literacy, and encouraged collaborative participation among students who often struggle with traditional text‑centered instructions. Students’ reflective writing further indicated an evolving awareness of how textual and visual modes interact, as well as a recognition of AI tools as interpretive mediators rather than neutral generators. Overall, the paper argues that integrating AI‑enhanced multimodal practices into literature teaching can broaden access to literary understanding, particularly in vocational contexts where experiential and visually oriented learning is pedagogically valuable. It also highlights the importance of human‑centered AI principles—transparency, critical engagement, and learner agency—in shaping meaningful educational uses of emerging technologies. By positioning students as active interpreters and evaluators of AI outputs, the intervention contributes to ongoing discussions on ethical and pedagogically grounded AI integration in education and illustrates the potential of text‑to‑image tools to enrich multimodal literacy in EPAL classrooms.
Beyond the Museum Walls: Policy proposals for the educational use of AI in disseminating national cultural heritage
(2025-12-12) Georgoula, Evangelia; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—Artificial Intelligence (AI) is emerging as a catalytic factor capable of radically transforming the educational utilization of National Cultural Heritage (NCH). It offers the potential to transcend the physical boundaries of museums, providing innovative methods for personalized access, dynamic interaction, and the integration of cultural content into the context of lifelong learning. This paper situates itself within the theoretical framework of Digital Transformation in the sectors of Culture and Education. It critically examines current practices and institutional challenges arising from the adoption of AI tools—such as recommender systems, virtual guides, and natural language processing—to enhance the educational dimension of NCH. The primary objective of this research is to formulate specific, actionable Policy Proposals that will facilitate the systematic and ethically responsible integration of AI into heritage education. Based on a qualitative review of international literature and a comparative analysis of European strategies, the study identifies a critical "institutional lag" between technological advancement and educational adaptation. The paper proposes a comprehensive policy framework focusing on establishing a National Strategy through inter-ministerial cooperation, designing mandatory professional development for educators, creating "AI-ready" digital repositories, and instituting robust ethical guidelines. Ultimately, the transition "Beyond the Museum Walls" is presented not merely as a technological upgrade, but as a critical political choice essential for fostering an inclusive and democratic cultural education.
The Peripheral Thought Accelerator: A Critical Pedagogical Framework for Integrating LLMs in Folk Guitar Education
(2025-12-12) Athanasiou, Ioannis; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The recent institutionalization of the Folk Guitar (Laikí Kithára) in the Greek music education system (Government Gazette 77/2022) introduces a pedagogical paradox: the necessity to codify an orally transmitted tradition into a rigorous diploma curriculum without compromising its embodied, improvisational essence. This paper proposes the "Peripheral Thought Accelerator" model, a critical pedagogical framework where Large Language Models (LLMs) such as Google Gemini and NotebookLM function not as generative performers, but as "Socratic interlocutors." Through a qualitative action research study of three distinct student profiles—a novice adolescent, a technical professional, and a mature scholar—we demonstrate that AI integration varies significantly based on the learner's epistemological background. The findings suggest that when the "core" of musical action is strictly human-governed, AI can effectively manage the "peripheral" cognitive load. However, this utility is contingent on the student's maturity, acting as a broadening tool for novices while serving as a dialectical partner for experts.
The Educator as an AI-based Mediator in the Era of Education 5.0: Reducing Educational Inequalities
(2025-12-12) Kota, Eris; Nikolidakis, Symeon; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The advent of new technologies and their integration into educational settings have carried a great promise to transform teaching and learning, to foster creativity, to advance critical thinking, and to democratize knowledge. The introduction of Artificial Intelligence (AI) applications in education is seen by some as a blessing and by others as a curse. The concerns are intense, mainly due to the fact that minors are involved, whose personalities are shaped within the educational process. This does not mean that we should ban the introduction of AI tools in education, but that we must limit their application with the basic aim of improving the educational process without losing sight of its fundamental goal, which is to shape responsible and conscious citizens and cultivate ethical thinking. This requires the educator's role to become that of a coordinator and advisor and not that of the transmitter of knowledge, moving toward an era of future education in which teachers and AI technologies work together.
From Theory to Practice: Implementing Participatory Pedagogical Approaches to Support Students’ Understanding of Artificial Intelligence
(2025-12-12) Tsioupli, Evgenia; Chatziaggelaki, Katerina; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract—The rapid development of Artificial Intelligence (AI) and its increasing presence in educational contexts call for pedagogically grounded and critically informed approaches to its integration in schools. This paper presents a participatory, school-based research project conducted with upper secondary students, aiming to explore their perceptions, attitudes, and concerns regarding the role of AI in learning and everyday life. The study was designed not only as an empirical investigation but also as a pedagogical intervention that foregrounds student agency and inquiry-based learning. Students participated actively as co-researchers in the formulation of research questions, the design of an anonymous questionnaire, data collection, and the interpretation of findings. The theoretical framework draws on inquiry-based pedagogy, critical digital literacy, and sociocultural theories of learning, emphasizing the role of human mediation and ethical reflection in AI-enhanced educational environments. The findings indicate that students adopt a nuanced and critical stance toward AI. While they recognize its potential to support understanding, organization, and efficiency in learning, they also express concerns related to dependency, data privacy, creativity, and the erosion of personal effort. AI is predominantly perceived as a supplementary cognitive tool rather than a substitute for learning or teaching. Students further emphasize the irreplaceable role of teachers as guides, mentors, and ethical mediators in AI-mediated learning contexts. Beyond the empirical results, the study highlights the pedagogical value of participatory research. Through their involvement in the research process and the public presentation of findings at a student conference, students developed critical thinking, research literacy, communication skills, and ethical awareness. The paper argues that engaging students as active producers of knowledge offers a meaningful pathway for fostering responsible, reflective, and pedagogically sound engagement with AI in secondary education.
Literacies, Multiliteracies, New Literacies and Literacies in the AI Era. The Society in the Post-Digital Age
(2025-12-12) Nerantzis, Nikolaos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract — The historical evolution of literacy reflects a gradual move away from a narrowly functional definition-literacy as a technical skill (e.g., reading, writing, arithmetic)-toward a concept that is socially situated and ideologically charged (Tuominen et al., 2005; Sperling et al., 2024). As D. Barton, M. Hamilton, and R. Ivanič (2000) argue, there are multiple literacies connected to different domains of life, and these are shaped by social institutions and relations of power. Multiliteracies brought into sharper focus the role of cultural and linguistic diversity, multimodal forms of expression, otherness, and social contexts in shaping communication and the ways people make sense of-and ultimately experience-their world (Cope & Kalantzis, 2000; New London Group, 1996). New literacies extended this perspective into digital, online, and collaborative modes of communication, suggesting that a literate citizen is defined through the social practices, values, and identities (both offline and online) that accompany them (Jenkins et al., 2009; Lankshear & Knobel, 2011). In today’s social landscape, the ongoing expansion of what counts as AI literacy brings an old question back with renewed urgency: what does it mean to be “literate” in a society where knowledge, judgment, and decision-making are increasingly permeated by big data, statistical inference, and algorithmic systems (Chiu et al., 2025)? From this vantage point, sociocritical analyses of contemporary literacies increasingly attend to three interrelated shifts: (i) a movement from the dominance of discourse to the mediation of data; (ii) a movement from citizen participation in content production-across both physical and digital worlds-to the co-production of meaning with machines; and (iii) a movement from critically reading social structures to critically reading algorithmic structures (Rapanta et al., 2025; Ruiz et al., 2025; Williamson & Eynon, 2024). The evolution of literacies in the age of AI creates significant opportunities, but it also introduces complex challenges. Societies do not always manage to absorb these ongoing shifts in an even or coherent way, partly because changes in communication and knowledge practices are rapid, multi-layered, and often asymmetric. The result is a set of educational, social, and digital divides-serious, but not inevitable or insurmountable. What becomes necessary, then, is the development of new capacities and skills: not only so that citizens can use emerging tools effectively, but also so that they can understand them in depth, question them, and help shape the norms that govern their use. As C. Rapanta and colleagues (2025) emphasise, literacies such as critical GenAI literacy should be understood “as a constellation of situated literacies, shaped by disciplinary perspectives, socio-political contexts, and technological affordances”. In a post-digital era-where AI technologies and digital tools are organically intertwined with everyday material and social practices-the challenge goes beyond simply reading the world of AI. It is also about rewriting it, and redesigning it critically and collectively.
Educators’ Digital Competencies: From Theory to Practice
(2025-12-12) Chatzikyrkou, Maria; Fountoukidou, Evanthia; Tsagaris, Apostolos; Πολυτεχνική Σχολή, Τμήμα Μηχανικών Πληροφορικής και Ηλεκτρονικών Συστημάτων
Abstract —Living in the 21st century, digital skills constitute a fundamental prerequisite for educators to create a learning environment that meets their students’ educational needs. In a world overwhelmed by technological challenges, teachers are called upon to apply ICT (Information and Communication Technologies) in practice during the learning process, becoming creators of digital content, facilitators of learning, and guides for their students. However, for teachers to be able to effectively implement educational technology in practice, continuous training, collaboration, experimentation, and time are factors that are all required. The result, nevertheless, is impressive, as the use of digital tools transforms teaching into a flexible, meaningful, and more participatory process with new perspectives that foster learning motivation. Technology thus becomes a valuable ally for the teacher, and theory finds its application in practice. The sample was formed using purposive sampling. The questionnaire was completed by teachers from primary and secondary education, regardless of specialty, who were participating in four different postgraduate programs delivered through distance learning. Structurally, the questionnaire is divided into two parts. The first part includes the demographic characteristics of the participants, which also define the independent variables of the study. These include gender, age, years of service, subject specialization, educational level, school structure, school location, level of studies, ICT training, and pedagogical education. The second part of the questionnaire consists of two main categories. The first category investigates participants’ accessibility to digital resources, while the second examines the obstacles to the effective integration of new technologies in education. Based on these dependent variables, the study initially explores and evaluates the academic and instructional background of teachers in relation to their level of access to digital resources and the barriers to the effective integration of new technologies in education. The questionnaire includes five Likert-scale subscales, with total scores calculated based on the mean responses for each category. The “Accessibility to Digital Resources” scale yielded an average score of 3.76 (SD = 0.882), while the “Barriers to the Effective Integration of New Technologies in Education” scale showed a mean score of 3.43 (SD = 0.578). Overall, the analyses indicate that accessibility to digital resources and perceptions of barriers to technology integration vary according to teachers’ demographic and professional characteristics. Regarding specialization, teachers in the Humanities and Technological Studies reported higher accessibility to digital resources compared to teachers in the Natural Sciences, though the differences were not statistically significant. A similar trend was observed for the barriers, as there were no statistically significant differences between teachers of different specialties. In terms of workplace location, teachers working in island and rural areas reported higher accessibility to digital resources, whereas teachers in urban areas displayed lower values, showing a statistically significant difference. Years of experience and educational level were also correlated with digital accessibility, with mid-career teachers and doctoral degree holders presenting higher scores. Finally, pedagogical training and ICT education were strongly associated with greater accessibility, particularly among SELETE/ASPETE-trained teachers and those holding Level B ICT certification. However, perceived barriers showed less variability and did not reach statistical significance. In conclusion, the study demonstrates that specialized education and professional experience enhance the use of digital resources in education, while barriers to successful integration appear to be less affected by demographic and professional factors.

Περιηγούμαι

Πρόσφατες Υποβολές