In 2024, I submitted a video to the UNSW Scientia Education Academy Exemplary Teaching Practice Awards (category: "Using AI in teaching and assessment") describing how I integrated Large Language Models into two Australian War College courses - ZEIT8205 Fundamentals of Surveillance Technologies and ZEIT8213 Communication and Information Systems.

The challenge was specific: my students are mid-career military officers from army, navy, and air force backgrounds. They are highly motivated but the majority had no programming experience. While the courses do not require coding, the ability to experiment with concepts in MATLAB - computing radar range equations, simulating radar systems, visualising signal processing - would substantially deepen their understanding.

I structured the integration across three assessment types:

Homework assignments (Design-3): Across 10 fortnightly assignments, students used LLMs to write MATLAB code that solved course-specific problems. A brief MATLAB introduction in the first weeks gave students enough foundation to evaluate and modify LLM-generated code rather than accepting it blindly.

Discussion forum (Design-2, Design-4): Students selected a radar or surveillance system, prompted an LLM to simulate it in MATLAB, and posted their prompts and results to a forum. Peers then replicated the submission using a different LLM and suggested improvements. This created a structured peer feedback loop where students critically evaluated both the AI output and each other's approaches.

Presentations (Design-1): Students selected topics relevant to their future military roles and used their acquired LLM skills to build working demonstrations. One team with no prior coding experience implemented a hypersonic glide vehicle simulator with a classification algorithm. Another team trained an AI model to detect and classify warships from imagery.

The results exceeded my expectations, officers who had never written code were building functional simulations within weeks. The approach treats AI as a tool that amplifies learning rather than replacing it.