The IDEMS Podcast

Michele and David explore two different approaches to building AI systems: tools that enable experts to work faster and more effectively, and tools designed to help more people develop and apply expertise. Drawing on Michele’s work building multi-agent systems for STACK, they discuss how AI can remove technical barriers while keeping human knowledge, judgement and meaningful contribution at the centre.

Lily and David reflect on a new world map that reveals just how much larger Africa is than conventional maps make it appear. They explore why familiar maps distort the size of countries and continents, the mathematics behind different map projections, and why representing Africa at its true scale matters for how we see and understand the continent.

Prompted by the recent potential AI-generated proof related to the Navier–Stokes problem, George and David reflect on what this could mean for mathematics and mathematicians. They explore the role of proof in advancing mathematical knowledge, what might change if AI becomes increasingly capable of producing proofs, and why mathematics has always been about more than solving problems alone.

Following their reflections on climate data rescue, James and David explore how the challenges of making rescued data genuinely usable are playing out in Zambia and Zimbabwe. Through their experiences working with meteorological services in both countries, they discuss the work needed to improve data quality, bring valuable historical records back into use, and build sustainable processes for keeping them usable over time.

James and David reflect on their experiences supporting climate data rescue across several countries. They discuss why preserving historical climate records is increasingly important for understanding a changing climate, and why rescuing data is only truly valuable when it can be put to use, exploring what it takes to bridge the gap between data rescue and real-world applications.

Continuing their reflections on 20 years of research methods support, Lily and Roger explore the often-overlooked problem of zeros in real data. Through examples from agricultural research, they discuss why understanding what zeros represent can change how data is summarised and reveal important patterns that might otherwise be missed.

Santiago and David discuss recent progress towards integrating STACK assessment into IDEMS’ open statistics textbooks. They explore how AI is helping match existing questions to textbook content, the work involved in adapting and improving those questions, and how richer forms of assessment can help students develop a deeper understanding of statistical concepts.

Continuing their reflections on 20 years of research methods support, David and Lily explore multilevel data through examples from agricultural research in West Africa. They discuss how identifying the right levels can help account for variability and reveal important differences that might otherwise be missed.

Continuing their reflections on 20 years of research methods support, David and Lily explore statistical interactions through a striking example from agricultural research in West Africa. They discuss how interactions can reveal the importance of context, and how understanding local practices can challenge research assumptions and lead to more relevant research.

David and Lily reflect on how the Explore, Describe, Present framing emerged through years of research methods support with the West African Community of Practice. They discuss how long-term collaboration helped develop a simple language for thinking about the different roles of data visualisation, including the importance of understanding your audience when presenting results.

Following the first STACK Conference to be held in Africa, David and Kate reflect on what the event revealed about IDEMS’ vision for open textbooks. They discuss growing interest in adapting the approach across different African contexts, and explore the importance of local ownership, collaboration, and working with existing education systems to bring open resources to scale. They also consider how bringing together experiences from high- and low-resource settings can spark new ideas and create opportunities for innovation that flow in both directions.

Following the International Conference on Teaching Statistics (ICOTS), David and Kate reflect on how IDEMS’ vision for open textbooks was received by the statistics education community. They explore how rapid developments in AI are changing what students need to learn, and why adaptable resources, carefully designed assessment, and meaningful human decision-making are becoming increasingly important. They also discuss the encouraging response from textbook authors, educators, and international organisations, and what this could mean for building a wider community around the open textbook initiative.

David and Mike reflect on a period of change and opportunity for mathematics education in Kenya. They discuss the upcoming launch of a new MSc in Mathematical Innovation, lessons from bringing open textbooks into classrooms, and the importance of putting teachers at the centre of educational innovation. Recorded just before the first STACK Conference to be held in Africa, their conversation also explores how bringing together local stakeholders and an international community could contribute to a more relevant, accessible, and locally grounded approach to mathematics education.

In the final episode of this three-part series reflecting on the International Conference on Teaching Statistics (ICOTS), David and Lily discuss the reception of IDEMS’ open textbook initiative within the statistics education community. They explore how new approaches to adaptation, accreditation, and collaboration could strengthen open educational resources, enabling communities to build on one another’s work while preserving the expertise and contributions of original authors.

In the second of three episodes reflecting on the International Conference on Teaching Statistics (ICOTS), David and Lily discuss the growing influence of AI on statistics education. They reflect on how the community is responding to rapid technological change, and explore why educators should play an active role in shaping AI tools that support responsible, accessible, and equitable learning.

In the first of three episodes reflecting on the International Conference on Teaching Statistics (ICOTS), David and Lily discuss the statistics education community itself. They explore why data literacy deserves to be recognised as a distinct educational priority, reflect on the longstanding relationship between statistics and mathematics, and consider why so many well-established ideas from statistics education research have yet to achieve broader impact.

David and Kate discuss the Farm Hack Box and the new possibilities it suggests for community-owned digital infrastructure. While conceived for farming communities, they explore how the same ideas could reshape education and other social impact contexts, enabling local AI, greater digital sovereignty, and technologies that communities can truly own and adapt.

David and Kate examine the idea of digital sovereignty through the lens of communities rather than nation states. Reflecting on recent discussions around AI governance and digital infrastructure, they explore why community ownership, local adaptation, and collaboration may be just as important as national strategies in shaping the future of technology.

David and Michele discuss how recent advances in AI are opening new possibilities for the Open App Builder and the authoring of digital tools for social impact. Drawing on IDEMS’ work with Parenting for Lifelong Health, they explore how AI agents can help reduce barriers to authoring while keeping people in control of the decisions that matter, enabling more adaptable and community-owned digital technologies.

David and Michele discuss how advances in AI are opening new possibilities for STACK question authoring. Using this practical example, they reflect on the design of multi-agent systems, the role of specialised models, and why building efficient, interoperable AI tools may be just as important as building ever more powerful ones.