Keynotes
Prof. Kate Borthwick
Professor of Digital Education in Languages, Cultures and Linguistics at the University of Southampton.
Prof. Borthwick is Lead for AI in Education at her university and founding chair of the University Digital Education Advisory Group. She is Director of the University open online course programme and an award-winning lecturer in applied linguistics and online learning design. She also serves as twice-elected Vice-President of the European Association for Computer Assisted Language Learning and is a member of the Editorial Board of ReCALL journal. Her research and teaching activities focus on digitally-mediated language learning and teaching, online education, and Artificial Intelligence in educational contexts.
Keynote: Rewriting the rules: the transformational power and challenge of AI in education
Use of generative AI has rapidly become a part of our daily life – in both our personal and professional lives as educators and researchers. We have all begun to live the incredible opportunities and possible risks that GenAI offers and the pace of technical development and change is so fast, it can feel like we’re ‘building the track while the train is running’. How do we balance the opportunities of AI with the risks? Does education change? Does the nature of being an educator change? What happens if we don’t like that change?
This talk will consider how educators and researchers might balance their concerns with their excitement about AI. It will consider the evolving challenges and opportunities that AI offers education and reflect upon the deepening nuances in opinions of AI in education from staff and students. It will draw upon my experience of leadership in AI in education at the University of Southampton and as an academic in digitally-mediated language learning and teaching, to offer ways to adopt or integrate AI into practice on your own terms as education experts, not on the terms of AI’s creators. It will advocate for human-centred, critical AI literacy as an imperative of education curricula.
Dr Velislava (Veli) Hillman
Founder of the EDDS Institute and Lecturer at Goldsmiths, University of London.
Dr. Hillman is an internationally recognised researcher specialising in the governance, safety, and real-world impacts of Artificial Intelligence and educational technologies. Through her work at the EDDS Institute, she investigates how data-driven technologies are shaping contemporary educational systems and the risks associated with insufficient regulatory frameworks.
Following a Visiting Fellowship at the London School of Economics (2020–2025), her research has focused on the impact of commercial technologies on children’s learning, teachers’ professional autonomy, and wellbeing.
She has advised major international organisations including UNESCO, the OECD, the World Bank, ChildFund Australia, and the Council of Europe, as well as schools and educational institutions worldwide.
Dr Hillman is also the author of Taming EdTech: Why Children Stand to Lose in an Unregulated Digitised Classroom (Bloomsbury Academic), a thought-provoking contribution to current debates on educational technologies, regulation, and the future of learning.
Keynote: AI Literacy for Soft Regulation – Who is Responsible for the AI Transformation?
What if AI literacy becomes the climate literacy of the AI age – teaching individuals how to adapt responsibly to a transformation whose structural causes they don’t control?
Artificial intelligence is entering education at extraordinary speed. But education is not experiencing AI in isolation. The same systems are simultaneously transforming labour markets, cultural production, knowledge infrastructures, public communication, finance, healthcare and military capabilities. The pace and scale of this transformation raise a question that is often missing from discussions of AI literacy: what exactly are we asking people to become literate in, and what responsibilities are we asking them to assume?
This keynote examines AI literacy as a form of soft regulation – a way of governing technological change through individual knowledge, responsible use, resilience, adaptation and self-regulation. AI literacy frameworks can certainly help people understand how AI works, recognise its limitations, verify outputs and use these systems more critically. But they cannot, by themselves, redistribute technological power, constrain corporate concentration, guarantee transparency, protect institutional autonomy or determine the social purposes for which AI should be deployed.
This creates an important parallel with environmental literacy. Over decades, citizens have been encouraged to recycle, reduce consumption, lower their individual carbon footprints and become more environmentally responsible. These practices can matter, but they do not substitute for regulating the infrastructures, production systems and corporations responsible for a substantial proportion of environmental harm. Academic institutions should therefore focus on asking whether we are beginning to reproduce the same model of responsibility!
As AI adoption accelerates across education, employment, culture and society, are we preparing people to shape these systems – or primarily to adapt to them?
Drawing on recent research into global AI literacy frameworks, the keynote will examine what current approaches emphasise and what they leave comparatively underdeveloped including questions of power, political economy, institutional sovereignty, epistemic authority, labour, environmental cost and democratic agency. It argues for moving beyond an AI literacy centred on technological competence and practical use towards a broader conception of literacy concerned with who controls AI, whose interests it serves, what infrastructures it depends upon, and what forms of human agency must remain possible in an AI-mediated society.
Prof. Francesco Peluso Cassese
Full Professor of Didactics and Special Pedagogy at Pegaso University and President of the University Quality Assurance Board.
Prof. Peluso Cassese is Editor-in-Chief of the Italian Journal of Health Education, Sports and Inclusive Didactics and has held several academic leadership positions, including Department Director and Coordinator of Ph.D. programmes.
He is actively involved in international research initiatives, including the Horizon Europe project ELOQUENCE, and serves on the Board of Directors of the Italian Society of Special Pedagogy (SIPeS).
His research focuses on educational neurosciences, inclusive didactics, and the pedagogical impact of digital transformation.
Keynote: What Kind of Educational Relationship Does a Large Language Model Construct? A Qualitative
Analysis of Human Annotations in University Guidance Dialogues
The study, was developed within the European ELOQUENCE project, starts from a question that I consider central as Large Language Models increasingly enter educational contexts: what kind of educational relationship does an Artificial Intelligence system construct when interacting with a student? More specifically, we addressed this question by analyzing university guidance dialogues generated through different Large Language Models. This work results from a collaboration between the research group of Pegaso Online University and the Italian National Research Council, CNR. The starting point of this research is a shift in perspective. When we evaluate a Large Language Model used in university guidance, we usually ask whether it provides accurate information, whether it shows bias, or whether it ensures fairness. We wanted to add another question: does the AI-generated conversation also have pedagogical quality? Through the analysis of human annotations, six dimensions emerged: pedagogical listening, recognition of student individuality, continuity, agency, reflective scaffolding, and educational reciprocity. The study is situated within the framework developed in the ELOQUENCE project and builds on previous work devoted to the generation of synthetic personas and the elicitation of bias.vThe process can be summarized in three main stages. First, synthetic student profiles with different characteristics are created. Then, these synthetic students interact with artificial counselors, generating guidance dialogues. Finally, human annotation is carried out to evaluate the dialogues. Our work introduces an additional conceptual step: annotations originally used to identify bias are reinterpreted as evidence of the pedagogical quality of the relationship. We therefore need to evaluate the system’s ability to maintain continuity, listen, personalize, promote autonomy, support reflection, and preserve appropriate educational reciprocity. From this perspective, human annotations also take on a different function. They are not only a tool for bias detection, but can also become a form of pedagogical evidence. This is precisely the main contribution we propose: integrating the evaluation of Artificial Intelligence with categories derived from educational theory.