How AI Learns: A Plain-Language Guide to Training, Data & Models

What you’ll explore

• Explain in plain language how AI models are trained on data, using the predict, check, adjust and repeat loop, and distinguish pre-training from fine-tuning
• Describe what tokens, parameters and context windows are, and why they matter for how an AI tool remembers, forgets and answers
• Explain why AI can confidently produce incorrect or out-of-date information (hallucinations and knowledge cut-offs) and check outputs in proportion to the stakes
• Describe how bias in training data reaches AI output and how human feedback (RLHF) shapes AI behaviour, and review people-facing output for fairness
• Use an understanding of training to get better results from AI tools, choose sensible ways to customise AI, and protect personal information when doing so

Who it’s for

Curious professionals who want to understand AI beyond the surface level Anyone who uses AI tools and wants to know why they behave the way they do Managers and decision-makers evaluating AI solutions Team members who have completed AI Fundamentals #01 to #03 and want to go deeper