Module 1
Move from structured interaction toward security-aware orchestration while preserving the distinction between literacies, collaboration levels, and AI fundamentals.
Participants move from vague asking to structured request design. The Friday should make visible how goal, context, constraints, source expectations, and response form change the conditions under which an LLM predicts and generates.
Participants learn to improve the input context before production begins. The Friday should show how clarification surfaces assumptions, constraints, audience needs, and success criteria, thereby changing downstream output quality.
Participants make model behaviour experimentally observable by holding a task stable while varying generation conditions. The Friday should connect output differences to tokens, attention, sampling, temperature, repeated runs, and model variability.
Participants learn that reliable output depends on the informational basis available to the model. The Friday should make grounding, retrieval, chunking, source selection, and evidence inspection observable.
Participants shift from single-pass generation to designed co-construction. The Friday should show how phases, roles, perspectives, intermediate artefacts, and checkpoints create control surfaces.
Participants learn to use AI not only for production but for critique, evaluation, and revision. The Friday should separate construction, judging, and improvement through explicit criteria and artefacts.
Participants learn that plausible AI output does not automatically create legitimate or accountable decisions. The Friday should focus on decision rights, escalation, oversight, contestability, evidence requirements, and responsibility.
Participants learn to design or tightly specify agentic workflows with state, tools, resources, prompts, permissions, validation, and control points. The Friday should connect orchestration to reliability, traceability, and accountability.
Participants learn to treat LLM security as a control problem in human-AI collaboration. The Friday should make visible how untrusted text, retrieved sources, tool permissions, validation steps, approval checkpoints, and human responsibility shape whether a workflow remains controllable.