Artificial intelligence is becoming remarkably capable of making predictions, generating content and carrying out increasingly complex tasks. But understanding what AI can do is only part of understanding how to use it well.
How do AI systems actually work? How can we establish whether their outputs are reliable? And what does it take to turn technical capability into useful outcomes and real-world impact?
Understanding AI for the Real World explores these questions in three parts.
Part I — Foundations: How AI Works
Understand the essential ideas behind machine learning, neural networks, deep learning, foundation models and generative AI, without requiring advanced mathematics.
Part II — Evidence and Reliability: From Data to Reliable AI
Explore how data are generated, how AI systems are tested, and why measurement, validation, generalisation and evidence matter when evaluating their claims.
Part III — Decisions and Impact: From AI Outputs to Real-World Impact
Examine how AI outputs inform human judgment and decisions, what happens when systems are implemented, and how we assess whether they create genuine value.
Written from the perspective of an engineer with nearly two decades of experience researching AI and its applications in healthcare, the book combines accessible technical explanations with scientific evidence and examples from healthcare, education, finance, public services and everyday life.
The aim is neither to celebrate AI uncritically nor to catalogue its failures, but to develop a clearer understanding of its capabilities, limitations and practical possibilities.
This is a book in progress. Chapters are being published progressively and can be read independently as the wider book develops. Reader questions, comments and constructive criticism are welcome.