Periodic Reporting for period 2 - augMENTOR (Augmented Intelligence for Pedagogically Sustained Training and Education)
Reporting period: 2024-01-01 to 2025-12-31
augMENTOR leverages the potential of Artificial Intelligence, particularly Explainable AI, and Knowledge Representation and Reasoning for instructional purposes and delivers an advanced pedagogical solution grounded on a pedagogical framework that bridges the gap between conceptual design and real-world implementation.
The project delivered a two-level pedagogical framework for implementing technology-augmented educational scenarios in diverse settings ranging from higher education to specialized adult training focusing particularly on the development and assessment of key 21st skills.
Our AI-boosted blended learning platform seamlessly integrates the pedagogical framework and utilizes a Knowledge Graph to unify data and deploy specialized algorithms for personalized learner profiling and actionable feedback, while strictly adhering to privacy-by-design and human-centric principles.
The augMENTOR solution was deployed and validated across diverse pilot settings representing various educational and training contexts. This validation process defined best practices for the ethical and effective use of augmented intelligence.
Ultimately, the augMENTOR platform acts as a collaborative partner that extends professional judgment, contributing to a resilient and sustainable digital society where technology serves as a catalyst for social inclusion.
The project delivered a two-level pedagogical framework for implementing technology-augmented educational scenarios in diverse settings ranging from higher education to specialized adult training focusing particularly on the development and assessment of key 21st skills.
Our AI-boosted blended learning platform seamlessly integrates the pedagogical framework and utilizes a Knowledge Graph to unify data and deploy specialized algorithms for personalized learner profiling and actionable feedback, while strictly adhering to privacy-by-design and human-centric principles.
The augMENTOR solution was deployed and validated across diverse pilot settings representing various educational and training contexts. This validation process defined best practices for the ethical and effective use of augmented intelligence.
Ultimately, the augMENTOR platform acts as a collaborative partner that extends professional judgment, contributing to a resilient and sustainable digital society where technology serves as a catalyst for social inclusion.
The augMENTOR project successfully created an AI-boosted educational ecosystem by combining advanced data science with robust pedagogical frameworks, following initial systematic literature reviews and user requirements elicitation.
Key technical achievements included the design and implementation of the augMENTOR reference architecture, a layered microservices architecture supported by a Knowledge Graph which seamlessly managed heterogeneous LMS data. Scientifically, the project advanced learner profiling using techniques like Gaussian Mixture Models to create dynamic, multidimensional profiles.
Pedagogically, the project delivered the validated augMENTOR Pedagogical Framework, comprising the macro-level Pedagogical Design Model with Emerging Technologies (PeDeMET) and the micro-level Technology-augmented Educational Scenarios and e-Activities (TESA), which incorporates Creative Pedagogy. This framework facilitated the creation of practical, technology-augmented scenarios, integrating an assessment strategy for 21st-century skills via explainable AI (XAI) recommendations.
Full-scale validation was achieved through diverse pilot implementations, confirming the versatility and effectiveness of the AI-boosted feedback mechanisms, with high educator utility and learner support. A "Human-in-the-Loop" philosophy ensured the AI complemented, rather than replaced, pedagogical judgment.
The project also established a rigorous legal and ethical framework, ensuring GDPR compliance and ethical oversight, supported by a dedicated Ethics Manager and an external Ethics Advisory Board, adhering to high European standards for trustworthy AI.
Finally, the project produced a set of policy briefs based on outcomes, lessons learned, and input from national and international policy-making events.
Key technical achievements included the design and implementation of the augMENTOR reference architecture, a layered microservices architecture supported by a Knowledge Graph which seamlessly managed heterogeneous LMS data. Scientifically, the project advanced learner profiling using techniques like Gaussian Mixture Models to create dynamic, multidimensional profiles.
Pedagogically, the project delivered the validated augMENTOR Pedagogical Framework, comprising the macro-level Pedagogical Design Model with Emerging Technologies (PeDeMET) and the micro-level Technology-augmented Educational Scenarios and e-Activities (TESA), which incorporates Creative Pedagogy. This framework facilitated the creation of practical, technology-augmented scenarios, integrating an assessment strategy for 21st-century skills via explainable AI (XAI) recommendations.
Full-scale validation was achieved through diverse pilot implementations, confirming the versatility and effectiveness of the AI-boosted feedback mechanisms, with high educator utility and learner support. A "Human-in-the-Loop" philosophy ensured the AI complemented, rather than replaced, pedagogical judgment.
The project also established a rigorous legal and ethical framework, ensuring GDPR compliance and ethical oversight, supported by a dedicated Ethics Manager and an external Ethics Advisory Board, adhering to high European standards for trustworthy AI.
Finally, the project produced a set of policy briefs based on outcomes, lessons learned, and input from national and international policy-making events.
The augMENTOR project has moved beyond the current state of the art by operationalizing a "human-centric" AI ecosystem that prioritizes pedagogical quality over simple automation. By the project’s conclusion, we have achieved the following key innovations:
Explainable and Transparent AI (XAI) was used to deliver a platform that goes beyond black-box AI to a transparent system. Using the augMENTOR Knowledge Graph, the platform provides natural-language explanations for every recommendation. This allows teachers and students to understand the reasoning behind the AI’s advice, ensuring technology serves as a reliable pedagogical companion.
The TESA & PeDeMET Frameworks designed offer a validated model for integrating AI into education. By grounding every insight in the TESA framework, the augMENTOR solution successfully transitioned from theoretical advocacy to a functional decision-support system, ensuring that digital tools are guided by proven teaching methods rather than technical capability alone.
The creative pedagogy framework designed, supports the empirical assessment of 21st-Century competencies. The project pioneered a methodology to measure the 4Cs (Creativity, Critical Thinking, Collaboration, and Communication). Utilizing finalized assessment rubrics and machine learning, we provided robust evidence of effectiveness across pilots. Results confirmed a significant increase in educators' confidence and a measurable improvement in learners' critical thinking and collaborative skills.
The augMENTOR platform, offers AI-driven evidence-based policy support through a dedicated ‘Policy maker’ view that interweaves Knowledge Graph data with augMENTOR Profiles to deliver policy recommendations to educational institutions and policy makers based on the outcomes of the pilot courses. This pipeline transforms complex learning data into structured strategic reports, enabling institutional leaders to identify learning gaps and make informed, human-centered decisions to improve inclusion and equity.
Appropriate standards for data safety were established supporting ethical "Privacy-by-Design", by keeping student identities local to their schools and providing fully anonymized datasets via Zenodo for scientific use, we proved that AI can be both powerful and strictly compliant with European privacy values.
Explainable and Transparent AI (XAI) was used to deliver a platform that goes beyond black-box AI to a transparent system. Using the augMENTOR Knowledge Graph, the platform provides natural-language explanations for every recommendation. This allows teachers and students to understand the reasoning behind the AI’s advice, ensuring technology serves as a reliable pedagogical companion.
The TESA & PeDeMET Frameworks designed offer a validated model for integrating AI into education. By grounding every insight in the TESA framework, the augMENTOR solution successfully transitioned from theoretical advocacy to a functional decision-support system, ensuring that digital tools are guided by proven teaching methods rather than technical capability alone.
The creative pedagogy framework designed, supports the empirical assessment of 21st-Century competencies. The project pioneered a methodology to measure the 4Cs (Creativity, Critical Thinking, Collaboration, and Communication). Utilizing finalized assessment rubrics and machine learning, we provided robust evidence of effectiveness across pilots. Results confirmed a significant increase in educators' confidence and a measurable improvement in learners' critical thinking and collaborative skills.
The augMENTOR platform, offers AI-driven evidence-based policy support through a dedicated ‘Policy maker’ view that interweaves Knowledge Graph data with augMENTOR Profiles to deliver policy recommendations to educational institutions and policy makers based on the outcomes of the pilot courses. This pipeline transforms complex learning data into structured strategic reports, enabling institutional leaders to identify learning gaps and make informed, human-centered decisions to improve inclusion and equity.
Appropriate standards for data safety were established supporting ethical "Privacy-by-Design", by keeping student identities local to their schools and providing fully anonymized datasets via Zenodo for scientific use, we proved that AI can be both powerful and strictly compliant with European privacy values.