by Zohre Mohammadi Zenouzagh, Leiden University, the Netherlands.
Artificial Intelligence (AI) is rapidly entering higher education, bringing with it a wave of tools that promise to personalise learning, automate routine tasks, and offer new ways to support students and educators. Yet as the excitement grows, so do the important questions: Which tools genuinely improve learning, and which simply add noise? Which align with our educational values? And how can we ensure that AI supports rather than replaces human educators?
At the Leiden Learning and Innovation Centre (LLInC), we believe that responsible use of AI in education starts with asking the right questions. That’s why our team is developing a practical evaluation framework for educational AI tools—one that goes beyond technological novelty and focuses on how these tools can foster thoughtful, engaged, and inclusive learning experiences. Our aim is to empower educators by providing clarity, confidence, and a shared language for evaluating what responsible and effective AI in education looks like.
This framework takes the form of a research-based rubric designed to guide real-world decision-making about AI use in teaching and learning. The theoretical foundation for this study was the Community of Inquiry (CoI) framework, which was selected for its relevance in assessing learning environments and pedagogical quality, particularly within blended and digitally mediated educational contexts. Originally developed by Garrison, Anderson, and Archer (2000), the CoI framework conceptualises meaningful learning as the intersection of three core dimensions: Cognitive Presence, Teaching Presence, and Social Presence. For this study, and in recognition of the ethical concerns raised in institutional documents about GenAI use in education, a fourth dimension—Ethical Presence—was added to the framework to reflect responsible and sustainable use of AI technologies in pedagogical settings. Guided by this model, we analysed a range of documents and research on AI in education to identify recurring themes such as guidance, critical thinking, collaboration, feedback, and ethical use. These insights were then organised within the CoI framework, allowing us to create a structured evaluation rubric grounded in pedagogical principles.
Now in the piloting stage, this LLInC innovation project brings together a multidisciplinary team exploring how AI tools contribute to responsible, human-centred learning environments. Beyond measuring technical performance, we are focusing on how these tools shape authentic teaching and learning practices. In this phase, the rubric plays a central role. Each scenario we design is aligned with specific indicators in the rubric to ensure it reflects pedagogical best practices and institutional ethical standards. For instance, when designing a scenario where an AI tool helps scaffold an argumentative essay, we examine whether it promotes higher-order thinking, aligns with learning outcomes, supports student voices, and ensures transparency and responsible AI use. At the same time, these scenarios serve as test cases for refining the rubric itself. By applying the rubric to evaluate how different AI tools perform across multiple use cases, we can assess its comprehensiveness and sensitivity. This iterative process helps ensure that the rubric becomes a reliable and practical instrument for evaluating AI in real educational contexts.
The insights gathered from educators are helping us continually refine our evaluation criteria and strengthen the connection between AI innovation and pedagogy. It is about using technology to deepen learning, support educators, and strengthen our shared responsibility for creating thoughtful and inclusive classrooms. Through this rubric, teachers can make informed, confident choices about integrating AI in ways that reflect sound pedagogy rather than hype. The result is not a prescriptive list of tools, but a “compass” that helps teachers navigate the complex landscape of AI, ensuring that human judgment remains at the heart of learning.
As we move forward, we remain committed to keeping educators at the centre of the conversation. Our goal is to move from AI hype to meaningful, evidence-based practice—ensuring that technology enhances, rather than overshadows, the human dimensions of education. We will share the results of our piloting in a follow-up article.
Would you like to learn more or get involved?
We welcome collaboration and dialogue as we continue to shape a responsible and inclusive future for AI in education. You can contact Z.mohammadi.zenouzagh@llinc.leidenuniv.nl.
Zohre Mohammadi Zenouzagh, Leiden learning and innovation centre, Leiden university, The Netherlands.
