Practical AI Fundamentals: LLM Behaviour, Prompts, and Risk Reduction
Location
In-person
Date & Time
Tuesday 28 Apr 2026 14:00 - 15:00
Audience exposure level: Open to all
This beginner-friendly course introduces the fundamentals of modern AI, with a focus on large language models (LLMs) and the practical factors that shape their outputs. Participants will place LLMs within the wider AI landscape, then develop a working understanding of how these models are trained, what training enables (and limits), and why they sometimes produce confident but incorrect answers. The course explores what influences model behaviour in practice, including differences between platforms, prompt design, and instructions that guide the model (system and custom instructions), as well as connecting an LLM to tools (for example, search, documents, or other software). Short examples are used throughout to show how these choices affect reliability and outcomes. The course also covers common risks—such as inaccuracy (hallucinations), bias, and misuse—and introduces practical techniques to reduce risk so these systems can be used responsibly in real settings.
Objectives
- Explain at a high level how LLMs are trained and why this creates common limitations
- Identify practical factors (platform settings, instructions, prompts, and tools) that change outputs
- Apply basic techniques to improve reliability (e.g., checking claims, structuring prompts, using sources)
- Recognise common failure modes (hallucinations, bias, unsafe outputs) and reduce risk in everyday use
- Describe where LLMs sit within the broader AI landscape and how they differ from other AI approach
