The IDEMS Podcast

Michele and David discuss the Turing test, and its relevance today. They explore various philosophical questions about intelligence, the limitations of the Turing test, and the ethical dilemmas posed by AI, particularly in the context of self-driving cars. David emphasises the vital role of human observation in the Turing test and expresses skepticism about society's ability to make responsible choices regarding AI regulation.

Michele and David discuss the impact of AI in low resource environments. They discuss the complexities surrounding AI technology, the hype versus the actual value, and the potential for AI to either widen or reduce global inequalities. They consider the need for robust infrastructural and social frameworks, the promise of small language models, and the importance of local ownership in AI development.

Lily and David explore a powerful framework for data analysis: Explore, Describe, Present. They discuss the importance of exploring data to understand its structure, describing data in the context of specific objectives, and effectively presenting insights to various audiences. Highlighting the challenges of modern data analysis, including the role of AI and the influence of tools like the tidyverse and R-Instat, they emphasise the need for structured approaches to make sense of complex datasets.

David talks to Rikin Gandhi from Digital Green to discuss the organisation's innovative approach to integrating AI with farmer support systems. They discuss Digital Green’s approach to working with AI, including the importance of human-in-the-loop systems, the benefits of using multimodal inputs like voice, text, and images, and the advantage of open-source data for tuning AI models to meet local agricultural needs. They also explore the potential and challenges of leveraging small language models to provide tailored support to farming communities and the critical role of local expertise in enhancing AI's effectiveness.

David and Kate delve into the ongoing AI boom, questioning whether it's mere hype or has real substance. They explore the ethical and responsible use of AI, emphasizing the importance of making technology accessible and beneficial to low-resource communities. They argue that small language models could provide specific, efficient solutions. The conversation also touches on the societal impacts of AI, the need for regulatory frameworks, and the potential for AI to democratize technology, moving away from its current gatekept state.

David and Mike discuss Kenya's new competency-based curriculum and a UK-backed campaign to create innovative digital textbooks. They delve into the challenges and potential solutions, highlighting the role of PreText and STACK technologies in revolutionizing education across Africa. https://www.crowdfunder.co.uk/p/open-digital-textbooks

In this episode, Michele and David discuss the development and impact of an AI tool for authoring STACK questions. They explore the potential of AI to enhance educational resources, make technology development more accessible, and address inequalities in low-resource environments. The conversation highlights both the opportunities and challenges presented by rapid advancements in AI.

Recently back from the World Statistics Congress, David and James reflect on the significance of this event, which brings together statisticians and the broader statistics community to share knowledge, foster collaborations, and advance the field. They discuss the evolving relationship between statistics and data science, highlighting the importance of both fields in the era of big data. They also explore the role of the International Association for Statistical Education (IASE) and the International Statistics Institute (ISI) in enhancing statistical education and integrating data science into curricula.

In this special 200th episode, Santiago and Johnny discuss how the podcast gets made. They consider the inspiration that led to its creation, the team members involved, and the evolution in production quality over the past 100 weeks. They also highlight the technical aspects, including the software and tools used, the editing process, and the role of AI in simplifying their workflow.

James Musyoka interviews David Stern, about his attendance at the Pan-African Convening on the Future of Biodigital Technologies in Food and Agriculture. Held in Addis Ababa, Ethiopia, and organized by the Alliance for Food Sovereignty in Africa (AFSA), the event focused on the impact of digital technologies on African agriculture and data sovereignty. David shares insights about a declaration formed during the event and discusses the three models of technology development presented by Million Belay. They draw parallels between these topics and IDEMS' African Data Initiative and the R-Instat project, highlighting the challenges and opportunities in achieving local ownership and...

David is joined by Michele Pancera, an Impact Activation Fellow at IDEMS. Michele shares his journey from a high school maths enthusiast to member of IDEMS, via being a teacher and beekeeper. The conversation covers Michele's passion for making mathematics engaging through games, his experience of societal perceptions of maths education, and his various professional pivots. The discussion emphasises the value of generalists in a world that often prioritises specialists, and touches on Michele's ongoing contributions at IDEMS including developing an AI assistant for STACK questions and his woodworking ventures.

Santiago and David explore the transformative potential of gamifying education. They discuss how rethinking exams to be more like driving tests could shift the educational landscape away from high-stakes differentiation to a mastery-based model.

David and Lucie discuss ongoing efforts to responsibly use AI to assist smallholder farmers in the Sahel region. They discuss the challenges faced by farmers in identifying pests and diseases and the shortcomings of commercial AI tools. They highlight the work of Digital Green, a non-profit organisation developing an AI tool for farmers that emphasises relevance, local language integration, and responsible AI practices. They explore the potential for collaboration and future advancements in AI tools tailored for low-resource environments.

Lily and David discuss the old adage “When all you have is a hammer, everything looks like a nail”, and how it applies to the realm of research, particularly in the context of statistical analysis and methodology. Emphasising the importance of having a diverse set of tools, they explore how narrow training in data skills can limit the effectiveness of research. They consider how methodologies from different disciplines can benefit from cross-disciplinary approaches, including medical science, climate, education and agriculture.

Wrapping up the series of episodes marking twenty years of research method support for the Collaboration for Resilient Food Systems, Lucie and David Stern consider the complexities of research planning. They explore the shifts in their roles towards structural change, particularly in the West African region. David shares his experiences with building local capacity and structurally embedding research methods within organisations like FUMA Gaskiya. They also reflect on the challenges of meeting researchers where they are and the importance of involving both qualitative and quantitative methods in research planning.

Lily and George discuss their personal experiences of using generative AI in their work. They explore how AI assists in course development, coding, and writing tasks, sharing insights on how these tools can enhance productivity and creativity.

Lucie and Roger discuss the intricacies and applications of crop simulation models in agricultural research. Roger explains the historical development of these models since the 1980s and provides a detailed explanation of how crop simulation models work, such as the DSAT and APSIM systems, which are used extensively in the United States and Australia. The discussion underscores the models' value in complementing traditional field experiments, especially in understanding long-term agricultural dynamics. They also touch on the challenges of implementing these models in Africa due to climatic data availability.

Santiago and David discuss how aspects of education could be “gamified” in order to improve outcomes. They explore various real-world examples of how games are used to enhance learning, including platforms like Kahoot and innovative projects like SmileyCoin. As well as the benefits, they consider the potential downsides, such as excessive screen time and the impact of dopamine release on student concentration.

As part of their continuing conversations around research methods for agriculture, Lucie and Roger discuss the utility of conducting research trials over multiple years. They consider the importance of understanding climate variability, and the value of pilot experiments as well as the implications of adapting research methods based on initial findings and the benefits of engaging farmers extensively in the research process.

How do we understand the material living conditions of people on Earth today? In this episode of the IDEMS podcast, Lily and David discuss the influential book “Factfulness” by Hans Rosling. They explore Rosling's key concepts about global development, particularly the categorisation of the world into four income levels instead of the binary “developing” and “developed” terms.

Social impact scientist and anthropologist Lucie Hazelgrove Planel joins Roger Stern to discuss the intricate process of designing agricultural experiments. Using a real-world example involving 10 maize varieties and a field with 12 plots, they explore the challenges of fitting theoretical models to practical scenarios.

How do you present research data in a compelling way? In this episode, Lucie and David discuss the intricacies of creating impactful visualisations for projects. They consider the distinctions between descriptive, exploratory, and presentation graphs. They emphasise the importance of understanding your audience, whether it's scientific, agricultural, or policy-making, and how to adapt visualisations accordingly to tell compelling stories.

In order to collect research data, we need to take measurements. As part of their continuing conversations, Lucie and statistician Roger consider this critical topic of measurements in agricultural research. They explore various types of measurements, such as context measurements, objective-related measurements, and those that help explain variability, using examples from agricultural trials and anthropological studies. The conversation highlights the nuanced nature of collecting quality data and calls for thoughtful planning and pilot studies.

Santiago and David discuss recent advancements and challenges in educational technology, looking at the integration of the STACK electronic assessment system with the PreTeXt tool for textbook authoring. Reflecting on a successful African STACK Conference held in Ethiopia, they consider the broader impact of these innovations on education in Ethiopia and potentially other African countries, emphasising the importance of making technology accessible at the student, lecturer, and institution levels.

In their continuing conversations on Research Methods for Agriculture, Lucie and Roger discuss the importance and application of controls in agricultural research. They consider the importance of controls in general, and consider examples from both agricultural and medical experiments, emphasising the importance of ethical considerations when using controls. Roger shares insights on how to make controls meaningful, how to adjust experiments based on intermediate findings, and the potential pitfalls of rigidly adhering to control rules.

AI tools can be helpful in analysing quantitative data, but their potential utility in qualitative analysis might be less obvious, or even concerning. In this episode of the IDEMS Podcast, data scientist Lily Clements and social impact scientist Lucie Hazelgrove Planel discuss the usage of generative AI in analysing qualitative data. They explore the ethical implications, the importance of context, and the challenge of addressing biases in AI-generated analyses, as well as the evolving intersection of AI with traditional qualitative research methods.

Lucie Hazelgrove Planel and Roger Stern consider the importance of multilevel data in agricultural research. They discuss the concept of measurement units at different levels, such as individual, household, community, and village, and the challenges faced when dealing with multilevel data in practice. Using an example of a student’s experiment on tadpoles, they illustrate key principles in determining appropriate experimental design and data analysis.

In this episode, Lily Clements and David Stern discuss the “Tricot” method for participatory breeding trials. Short for “Triadic Comparisons of Technologies”, Tricot involves farmers testing three crop varieties and ranking them based on qualitative measures. They reflect on a recent workshop aimed at simplifying this complex analysis using custom R packages and the R-Instat software.

Lucie and Roger continue their discussions of research methods for agriculture, this time focusing on on-farm trials. They consider the benefits and challenges of conducting research on farms versus research stations, emphasizing the importance of farmer involvement in the research process. They consider the innovative “Tricot” method, which tests multiple crop varieties with minimal control from researchers to increase real-world applicability.

In this episode, David and Santiago debate the 'Dead Internet' Theory, which claims that AI-generated content will dominate the internet, making it less reliable. David challenges this theory, emphasizing the need for digital literacy, responsible use of AI, and the complex nature of trust in institutions versus individuals. They also discuss the implications of misinformation and the importance of critical thinking in society.

In this episode, Lucie interviews Roger about essential aspects of agricultural statistical experiments. They discuss treatment, layout, and measurement, using an irrigation and maize variety case study. Roger emphasizes the importance of clear objectives and balancing statistical rigor with practical agricultural considerations.

statistical and mathematical modeling. They explore how each field approaches modeling, and touch on hybrid models that incorporate both statistical and mathematical elements, and the significance of uncertainty in modeling predictions.

In this episode, Lucie and Roger discuss planning agricultural research, highlighting on-station and on-farm trials. They emphasize integrating diverse data collection methods, farmer involvement, and balancing research design and farmer participation.

In this episode, Lily and George discuss their journeys from their PhD to IDEMS’ Postdoctoral Impact Activation Fellowship. Lily also describes her transition to a permanent role while George describes what he imagines his transition might be like. An informal chat that reveals two different enough yet similar experiences joining IDEMS.

In this episode, Lucie talks to Roger about his work in agricultural research methods. Roger shares his experience with CCRP, the shift from on-station to on-farm research, the role of farmers' organizations, and advancements in data collection and analysis.

In this episode, Lucie and David discuss the 20th anniversary of the Collaboration for Resilient Food Systems’ Community of Practice in West Africa, and in particular the evolution of its Research Methods Support. The conversation underscores the impact of patient, long-term approaches and local capacity building in Niger, Burkina Faso, and Mali, as an introduction to a series of upcoming episodes that will delve into the program's history and learnings.

In this episode, Lily and David discuss the concept of proxy variables and feedback loops. They explore the use of proxies when direct measurement is impractical, using examples from agriculture and education. The discussion includes the pitfalls of educational performance metrics and university rankings, demonstrating the complexities and potential misinterpretations involved in using proxy indicators.

Lucie and David discuss the recent progress and training in building a research methods support team in West Africa. The conversation highlights their efforts in providing various trainings, and the development of local teams capable of delivering these trainings. They also emphasize the importance of reflective learning reports, local capacity building, and moving towards more engaging pedagogical methods through a gradual, patient approach, aiming for sustained growth and effective collaboration.

In this episode, George and Kate discuss the challenge of transforming deep R&D into tangible products. They explore the balance between simplicity and complexity, the need for early market validation, and the difficulty of securing private investment for ambitious long-term goals. The conversation touches on the unique product and funding strategies for social enterprises and highlights how IDEMS navigates these challenges while emphasizing the importance of impact over profit.

In this episode Johnny and Kate discuss the intricacies of marketing IDEMS, emphasizing the challenges of presenting its diverse and impactful work. They highlight the importance of a well-structured website, balancing in-depth technical information with user-friendly content, and the need to showcase successes without overshadowing their collaborative nature. They also touch on the evolution of IDEMS’ marketing strategy, the role of design, and the aim to create a cohesive narrative that resonates with both experts and laypersons.