In recent times, there have been some specific initiatives that recognize the importance of learner agency, such as OECD’s Learning Compass, Cambridge English’s Learning Oriented Approach, and the CEFR’s ELP. Despite this activity, in a recent open call to applied linguists inviting papers that discuss learner agency, none of the featured subtopics (e.g., agency in writing, agency in curriculum, agency in pedagogy, etc.), concern language assessment.

In this talk, I discuss the fact that learner agency and language assessment are not incompatible. Moreover, cultivating learner agency is critically important in today’s world, where learners need to be prepared not only for current times, but for adapting to rapidly-changing circumstances.

One issue, though, that language assessment specialists have to confront is that “agency” is complex and challenging to implement. It has simply been defined as “the capacity to act in the world,” but this definition is about the potential of agency. It does not provide guidance for how to implement it in language assessment.

Furthermore, often agency is regarded as an individual phenomenon, when in fact, agency cannot be adequately conceived of as an individual attribute. The socio-cultural context within which it is enacted needs to be considered as well as the ways in which individuals relate to the affordances in the context.

In addition, it is important to acknowledge that agency is neither static nor inherent in language learners. It can be learned, but it cannot be conferred, i.e., it is not about empowering learners. Rather, “enabling conditions” can be established to encourage its development. Finally, I suggest ways that the dynamic and relational process that characterizes learner agency can be incorporated into assessments, without at the same time tacitly endorsing a neoliberal perspective.

Diane Larsen-Freeman is Professor Emerita of Education and Linguistics, Research Scientist Emerita, and former Director of the English Language Institute at the University of Michigan. She is also Professor Emerita at the SIT Graduate Institute in Vermont and a Visiting Faculty Member at the University of Pennsylvania. Some of her books include Complex Systems and Applied Linguistics (2008, with L. Cameron), winner of the MLA’s Kenneth Mildenberger Book Prize, the third edition of Techniques and Principles (2011, with M. Anderson), the third edition of The Grammar Book: Form, Meaning, and Use for English Language Teachers (2015, with M. Celce-Murcia), and Second Language Development: Ever Expanding (2018). She edited the journal Language Learning for five years, and later chaired its Board of Directors. Among her other honors, she received the American Association for Applied Linguistics’ Distinguished Scholarship and Service Award, and last year she was inducted as a Fellow of the Cognitive Science Society.

Large-scale educational assessments have long been framed as objective measures of learning, with their validity upheld as the gold standard of fairness and rigor. But what happens when the very systems meant to ensure validity perpetuate harm—particularly for multilingual, racialized, and other marginalized learners? In this keynote, I interrogate the field’s dominant approaches to validation and argue that traditional validity frameworks often obscure the violence they help maintain: the erasure of cultural knowledge, the pathologizing of difference, and the routine sorting of students into deficit-based categories. Drawing on justice-oriented scholarship and classroom-based examples, I propose a new path forward—one where validity is not just about evidence and interpretation, but about repair, recognition, and responsibility. Justice should not simply be an add-on to technical quality; it should be the foundation. If assessment is to serve all learners, our validity must begin—and end—with justice.

Dr. Jennifer Randall is the Dunn Family Professor of Psychometrics and Test Development in the School of Education at the University of Michigan. She is also the founder and Chief Visionary of the Center for Measurement Justice, which is a research center dedicated to facilitating the adoption and implementation of justice-oriented educational assessment and measurement processes. She received her Bachelor’s (1996) and Master’s (1999) degrees from Duke University and her doctoral degree from Emory University (2007). She began her career as a high school public-school teacher in social studies working with racially and ethnically minoritized students. It was in this capacity that she began to recognize the ways in which traditional assessment practices cause deep and irreparable harm to the most marginalized students – the students the system should be seeking to serve the most. Her work seeks to disrupt white supremacist, racist logics in assessment through practices that prioritise justice and are explicitly and unapologetically antiracist.

Abstract coming soon

Gino Roncaglia is a Full Professor of Digital Publishing, Digital Humanities, and the Philosophy of Information at Roma Tre University. He is a member of the scientific council of the Fondazione Scuola dei beni e delle attività culturali, a member of the board of directors of Biblioteche di Roma, and an Honorary Professor at the Universidad Nacional de Villa María in Argentina. He has been collaborating with RAI Cultura for over 25 years as an author and scientific consultant for many programs dedicated to the world of the Internet and network cultures; he has also published several books and numerous scientific articles on these topics. His most recent books include: "The Architect and the Oracle. Digital Forms of Knowledge from Wikipedia to Chat GPT" (Laterza 2023, 5th reprint 2024) and "Philosophy of AI" (Corriere della Sera editions, 2025). Before taking up a post at university, he was a documentalist, researcher and librarian at the Italian parliament.