Teaching soil as a system: interactivity, place, plural knowledge, and AI

<jats:p>Abstract. Reasoning about soil as a coupled system is a central aim of graduate training in soil science. Soil science is now often taught within broader degrees, without the specialists who once made systems teaching natural. Curriculum compression and lacking interactivity have kept systems reasoning out of many programmes. Large language models, now accessible to students, have made that absence harder to ignore. Here, we set out the design principles for an undergraduate soil science unit built around four requirements: interactivity, place, plural knowledge, and AI discernment. Interactive models, repeated landscape comparisons, Indigenous and Two-Way knowledge, and AI interrogation operationalise these requirements. Assessment, designed from intended outcomes, integrates them into one capability: place-specific reasoning from evidence. We provide a worked example that other soil science programmes can adapt.</jats:p>