Home | ‘Gender, accountability and curriculum policy – Reflections on a project using AI’ (2025), presentation at the 18th UKFIET conference on Education and Development, Oxford, 16th September 2025.

‘Gender, accountability and curriculum policy – Reflections on a project using AI’ (2025), presentation at the 18th UKFIET conference on Education and Development, Oxford, 16th September 2025.

Paper presentation: Gender, accountability and curriculum policy – Reflections on a project using AI, at the UKFIET Conference 2025, Mobilising knowledge, partnerships, and innovations for sustainable development through education and training. University of Oxford Examination Schools, Oxford, UK, 16 – 18 September 2025.

  • Ellen Weavers (UCL)
  • Eliza Ngutuku (UCL)
  • Elaine Unterhalter (UCL)
  • Helen Longlands (UCL)
  • Rosie Peppin Vaughan (UCL)
  • Mercy Mwongeli (UNESCO)
  • Caine Rolleston (UCL)
  • Esthery Kunkwenzu (University of Malawi)
  • Cresti Fitriana (UNESCO)

Abstract

SDG 4.7 focuses on all learners acquiring knowledge and skills to promote sustainable development, including sustainable lifestyles, human rights, and gender equality. However, identifying a robust indicator for this has been difficult. This paper reports on the Accountability for Gender Equality in Education (AGEE) project’s work to develop novel Natural Language Processing (NLP)/Large Language Model (LLM) protocols to analyse and identify concepts related to gender within curricula from Kenya, Malawi and Indonesia. This work represents a pilot for a larger initiative reviewing school learning materials.

The aspiration to review curricula and their contribution to gender transformative education was outlined in the Call to Action from the United Nations Transforming Education Summit (TES) of 2022. In 2023 a global accountability dashboard
to monitor progress against governments’ commitments was established. The dashboard reports on gender parity in enrolment, progression and attainment internationally and presents traffic light coding for key gender-related policies. However, an indicator is yet to be developed in relation to gender representation in curricula.

Academic literature on the use of AI to analyse gender representation within curricula remains scarce. The studies that exist (including textbook analysis) tend to rely on relatively simple NLP techniques reporting on a single aspect of the many ways in which gender can be represented in such documents. The paper reports on the participatory development of NLP/LLM techniques in relation to a rich Gender Framework for Curriculum Analysis that identifies wide-ranging ways in which gender can be represented in curricula. We reflect on the process of discussing the use of this protocol within the AGEE project in Malawi, Kenya and Indonesia The study explores emergent ethical and methodological issues, and the implications of this pilot for contributing transparent data on curricula to the TES dashboard and work in Malawi, Kenya and Indonesia on national AGEE dashboards.

For further information please see the AGEE working paper, ‘Versatile and expansive or biased and overcomplicated? Reflections on using Large Language Models (LLMs) to review ideas about gender in curriculum and policy documents‘.