Openness and trust in digital learning systems

🔍 European Digital Education Hub: Explainable AI (XAI) in Education This comprehensive report, developed by an expert group of practitioners, explores how explainability in AI is critical to fostering trust, accountability, and transparency in educational settings. 📘 What’s it about? The report provides guidance for policymakers, educators, developers, and institutional leaders on how to responsibly integrate explainable AI into teaching, learning, and governance systems. It links legal obligations (AI Act, GDPR) with ethical and pedagogical concerns. 📌 Key Themes and Insights: 🧠 Core Concepts of XAI ▪️Transparency, interpretability, explainability, and understandability are distinct but interconnected ▪️XAI bridges technical development and human comprehension, especially in complex, data-driven AI models ▪️ Transparency is not just a technical issue—it’s also an ethical and pedagogical imperative ⚙️ Legal and Ethical Foundations ▪️The EU AI Act does not mandate explainability outright but requires human oversight, fairness, and risk mitigation ▪️In education, this means AI systems must be understandable by learners, actionable for educators, and auditable by authorities ▪️ High-risk AI in education (e.g., grading, tutoring systems) must meet specific compliance criteria 👨🏫 Implications for Vocational Education and Training (VET) ▪️VET learners often engage with automated decision systems (e.g., intelligent tutoring, skill assessments) ▪️XAI ensures learners understand why an AI tool makes recommendations - vital for preserving learner agency, building metacognitive skills, and enabling self-regulated learning ▪️Teachers in VET must develop competences in AI literacy and critical thinking to use and explain these tools responsibly 🤝 Stakeholder Roles & Collaboration ▪️Developers must build for clarity and accountability ▪️Educators and institutions must scrutinise AI outputs, especially where decisions impact learners' futures ▪️ Multi-stakeholder collaboration is essential—particularly involving pedagogues in the design phase 📊 Pedagogical Design of Explanations ▪️Explanations should be tailored: local vs. global, simple vs. technical, and conditional vs. correlational ▪️In VET, actionable feedback (e.g., “you need to improve welding precision due to X pattern in your practice log”) is more effective than opaque scores 👩🏫 Educator Competences for XAI ▪️The report defines core digital and pedagogical competences for integrating XAI in curricula, including: ✔️Understanding AI models' logic ✔️Interpreting and communicating AI-driven outputs ✔️Teaching students to question, reflect, and act on AI advice ▪️Includes practical activities adaptable from for all leveles of education 📈 Conclusion Explainable AI isn’t a technical luxury—it’s a pedagogical necessity. In VET, it safeguards learner agency, fosters trust, and supports equitable learning outcomes #AIinEducation #DigitalCompetence Francisco Bellas European Commission

Yesterday, a student in my class candidly shared with me some of their go-to AI resources. That openness was a big moment for me—not because of the tools themselves, but because it showed me that they felt comfortable enough to talk freely about how they’re using AI in their work. It’s a sign that the trust we’ve been building in the classroom is paying off. When students start sharing how they’re leveraging AI without hesitation, you know the atmosphere you’ve created supports real learning and growth. Trust is the cornerstone of effective AI integration. Here are five ways I’ve worked to cultivate that trust: Be Transparent About AI’s Role: I’m upfront about how AI fits into our learning goals. I set clear guidelines but also explain the reasoning behind them, so students see AI as a supportive tool, not a replacement for their thinking. Show Vulnerability: I let students know that I’m also figuring things out as we go. By being honest about the learning curve I’m experiencing, I encourage them to be open about their own challenges and discoveries. Encourage Real-Time Conversations: When students mention how they’ve used AI, I don’t just nod and move on—I dive in. We talk through what worked, what didn’t, and how they approached it. This normalizes AI use and turns it into a shared learning experience. Celebrate Their Process: Whether they successfully apply AI or run into challenges, I make sure to recognize their efforts. This reinforces that AI is a tool for growth and experimentation, not just a quick fix. Model Responsible AI Use: I regularly demonstrate how I incorporate AI in my own work. When students see me using AI thoughtfully, they’re more likely to adopt similar practices, knowing that the tools have a real, practical role in our classroom. In the end, trust allows AI to become more than just another tool—it becomes part of a larger dialogue about learning, creativity, and innovation. And when students trust the process, they engage with AI more confidently and effectively. Amanda Bickerstaff Aco Momcilovic Brian Schoch Christina B. 👨🏫🤖 "Dr. Greg" Loughnane Goutham Kurra Iulia Nandrea Mike Kentz Michael Spencer Milly Snelling Anna Mills David H.

Deisigning for Education with Artificial Intelligence: An Essential Guide for Developers Executive Summary: This guide by the U.S. Department of Education provides essential insights and structured guidance for developers creating #AItechnologies in #education. It emphasizes responsible, safe, and equitable AI development aligned with #educationalgoals to build trust among #stakeholders in the #educationalecosystem, including #developers, #educators, and #students. Key Recommendations: 1. Designing for Education: Developers are advised to align their AI solutions with educational values, ensuring that AI supports human-centric, teaching, and learning goals. The guide stresses incorporating educators’ and students’ feedback throughout the design and implementation phases, promoting ethical, transparent, and inclusive AI. 2. Providing Evidence of Impact: To build trust, AI developers should base their designs on evidence of educational effectiveness and continuously monitor impact, especially for underserved groups. Clear documentation of rationale, impact assessment, and outcomes is recommended, with emphasis on disaggregated data to assess inclusivity and equity in performance. 3. Advancing Equity and Civil Rights: The guide calls for active steps to mitigate bias and algorithmic discrimination, ensuring that AI tools promote fairness and accessibility, particularly for marginalized and vulnerable groups. Developers are encouraged to understand and comply with federal civil rights regulations, embedding these principles into AI product development. 4. Ensuring Safety and Security: Safety in data handling, privacy protection, and security are emphasized as critical for AI in educational settings. The guide highlights the need for robust safeguards to protect users’ information and for transparent mechanisms that users can trust. 5. Building Transparency and Trust: The importance of open communication, clear reporting, and accountable practices is stressed to foster trust. Developers are urged to maintain transparency around their AI processes and be proactive in addressing potential risks and issues related to AI deployment. This guide underscores a collaborative approach, urging developers to work with educational stakeholders to ensure that AI technology is safe, ethical, and truly beneficial to diverse educational communities.

𝗣𝗿𝗲𝘀𝗲𝗿𝘃𝗶𝗻𝗴 𝗣𝗲𝗱𝗮𝗴𝗼𝗴𝗶𝗰𝗮𝗹 𝗧𝗿𝘂𝘀𝘁 𝗶𝘀 𝘁𝗵𝗲 𝗺𝗼𝘀𝘁 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁 𝗶𝘀𝘀𝘂𝗲 𝗶𝗻 𝗵𝗶𝗴𝗵𝗲𝗿 𝗲𝗱𝘂𝗰𝗮𝘁𝗶𝗼𝗻. I love Rachel Botsman’s notion of trust: “a confident relationship with the unknown.” 𝗣𝗲𝗱𝗮𝗴𝗼𝗴𝗶𝗰𝗮𝗹 𝘁𝗿𝘂𝘀𝘁 in teaching and learning is our collective capacity to stay in relationship amid uncertainty. AI brings with it ambiguity, tension, and constant change. Decisions will be contextual, choices often muddy, and clarity rarely permanent. The task is not to control uncertainty, but to navigate it together. My new article — 𝗙𝗶𝘃𝗲 𝗴𝘂𝗶𝗱𝗶𝗻𝗴 𝗽𝗿𝗶𝗻𝗰𝗶𝗽𝗹𝗲𝘀 𝗳𝗼𝗿 𝗻𝗮𝘃𝗶𝗴𝗮𝘁𝗶𝗻𝗴 𝗮𝗿𝘁𝗶𝗳𝗶𝗰𝗶𝗮𝗹 𝗶𝗻𝘁𝗲𝗹𝗹𝗶𝗴𝗲𝗻𝗰𝗲 𝗶𝗻 𝗦𝘁𝘂𝗱𝗲𝗻𝘁𝘀 𝗮𝘀 𝗣𝗮𝗿𝘁𝗻𝗲𝗿𝘀 𝗽𝗿𝗮𝗰𝘁𝗶𝗰𝗲 𝘁𝗼 𝗽𝗿𝗲𝘀𝗲𝗿𝘃𝗲 𝗣𝗲𝗱𝗮𝗴𝗼𝗴𝗶𝗰𝗮𝗹 𝗧𝗿𝘂𝘀𝘁 — offers a framework for doing precisely that. 1. 𝗖𝘂𝗹𝘁𝗶𝘃𝗮𝘁𝗲 𝗼𝗽𝗲𝗻 𝗮𝗻𝗱 𝗰𝘂𝗿𝗶𝗼𝘂𝘀 𝗰𝗼𝗻𝘃𝗲𝗿𝘀𝗮𝘁𝗶𝗼𝗻𝘀 𝗮𝗯𝗼𝘂𝘁 𝗔𝗜. You’ll find a guide and a set of prompt questions to ask yourself, to bring to your classroom, and to spark dialogue among colleagues. 2. 𝗔𝗰𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗮𝗻𝗱 𝘄𝗼𝗿𝗸 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗲𝗺𝗼𝘁𝗶𝗼𝗻𝗮𝗹 𝗮𝗻𝗱 𝘃𝘂𝗹𝗻𝗲𝗿𝗮𝗯𝗹𝗲 𝗹𝗶𝘃𝗲𝗱 𝗲𝘅𝗽𝗲𝗿𝗶𝗲𝗻𝗰𝗲𝘀 𝗼𝗳 𝗔𝗜 𝘂𝘀𝗲. Each principle includes reflective questions that help surface the emotions — curiosity, anxiety, frustration, hope — that accompany learning with AI. 3. 𝗠𝗮𝗸𝗲 𝘁𝗵𝗲 𝗿𝗼𝗹𝗲 𝗼𝗳 𝗔𝗜 𝘃𝗶𝘀𝗶𝗯𝗹𝗲 𝗮𝗻𝗱 𝗻𝗲𝗴𝗼𝘁𝗶𝗮𝗯𝗹𝗲 𝗶𝗻 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴. These questions invite transparency and collective negotiation of how AI enters teaching, assessment, and authorship. 4. 𝗖𝗿𝗲𝗮𝘁𝗲 𝗰𝗼𝗻𝘀𝗶𝘀𝘁𝗲𝗻𝘁 𝗼𝗽𝗽𝗼𝗿𝘁𝘂𝗻𝗶𝘁𝗶𝗲𝘀 𝗳𝗼𝗿 𝘀𝗵𝗮𝗿𝗲𝗱 𝗹𝗲𝗮𝗿𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗿𝗲𝗳𝗹𝗲𝗰𝘁𝗶𝗼𝗻 𝗼𝗻 𝗔𝗜. The prompts support building rhythms of reflection where students and staff learn together, acknowledging that understanding AI is an ongoing process, not a one-off conversation. 5. 𝗚𝗿𝗼𝘂𝗻𝗱 𝗲𝘁𝗵𝗶𝗰𝗮𝗹 𝗻𝗼𝗿𝗺𝘀 𝗶𝗻 𝘀𝗵𝗮𝗿𝗲𝗱 𝘃𝗮𝗹𝘂𝗲𝘀 𝗮𝗻𝗱 𝗮 𝗰𝗼𝗹𝗹𝗲𝗰𝘁𝗶𝘃𝗲 𝗺𝗼𝗿𝗮𝗹 𝗰𝗼𝗺𝗽𝗮𝘀𝘀. Here, the questions help communities articulate and live by shared principles of justice, care, and respect. I share this article in the hope that it offers both language and practice for educators and students who want to face uncertainty together — to co-design, co-create, co-discover, and co-inquire in ways that uphold pedagogical trust. 𝗕𝗲𝗰𝗮𝘂𝘀𝗲 𝘁𝗵𝗲 ‘𝗵𝗼𝘄’ 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 𝗺𝗼𝘀𝘁. We can have every AI framework, but none of it will work if we fail to preserve trust — the fragile, relational core that makes higher education what it is. Read more: https://lnkd.in/gaZsAMS3 International Journal of Students as Partners (IJSaP) thanks for feedback Mick Healey, Alison Cook-Sather, Meng Zhang, Yifei Liang, PhD. Tim Fawns, Danny Liu, Margaret Bearman, Christine Slade, Simon Buckingham Shum, Kylie Readman, Jason M. Lodge, Peter Felten I think you’ll agree.

📢 New EU Guide: Responsible AI in Education 🤖📘 The EducationalAI project has just released its Ethical & Legal Guide (Aug 2025) — a must-read for leaders and educators navigating the human-centered, responsible use of AI in classrooms and institutions. 🌍✨ Highlights: 💡 Who it’s for: Leaders, teachers, and administrators implementing AI in learning and operations 🏫📚 🔍 Ethical foundations: ✅ Human-centered: AI empowers educators, doesn’t replace them 🙋♀️🤝 ✅ Fairness & inclusion: Equal access, prevent bias, support diversity 🌈 ✅ Human oversight: Key decisions stay with people, not algorithms 👥 ✅ Transparency: AI must be explainable and understandable for all ages 🔍 ✅ Privacy: GDPR-aligned data protection and consent safeguards 🔐 ✅ Academic integrity: Promote honesty and clear AI-use policies ✏️📖 ⚖️ Legal framework: 📜 EU AI Act (2024/1689): Standards for safety, transparency, and risk levels ⚠️ 👶 Children’s rights: Ethical safeguards central to all AI-driven learning environments 💡 🔐 Data protection: Responsible AI use for children and vulnerable groups 🌱 Bottom line: AI in education should enhance learning, equity, and trust, not compromise them. This guide provides a roadmap for schools and universities to balance ethics, compliance, and innovation 🚀 #AIinEducation #ResponsibleAI #AIethics #HigherEd #EdTech #DigitalLearning #HumanCenteredAI #EducationalLeadership #FutureOfLearning #EthicalInnovation #DigitalPedagogy #AIIntegration #LearningTransformation #CriticalThinking #AIcompetencies #EducationFutures #PedagogyFirst 🌟 Follow and Connect: Woongsik Dr. Su, MBA