A pilot of the “AI Agents for Defence Education” workshop took place from 21 to 25 September 2026. Designed for defence educators, instructors and curriculum developers, the workshop focused on using AI agents to support learning activities under teacher supervision. Its approach connected practical experimentation with clear educational purposes and critical review.
The five-session programme progressed from agent fundamentals and technical setup to the design of a small educational prototype. Prepared starter projects and guided examples provided an entry point for educators without programming experience. The programme introduced AI-assisted coding alongside basic project review, while keeping the emphasis on instructional decisions.
The practical task centred on a bounded learning activity: defining a learning objective and learner task, assigning the agent a role, and writing instructions. The intended prototype also included a limited source or tool, test examples, and a short explanation of how the teacher would supervise its use.
Testing was a key part of the design. The programme addressed routine questions, ambiguous requests and situations in which the available evidence was insufficient. These cases provided a basis for reviewing agent behaviour, refining instructions and recognising when an AI response required further scrutiny.
Responsible use underpinned the pilot, including work with public, synthetic or institution-approved materials. The closing session focused on prototype demonstrations, limitations, and next steps. Further classroom testing and feedback remained necessary before concluding the effects on learning.