Intelligence, Data and Judgement

A textbook written to be taught from. Twenty-two chapters in five parts — foundations, the material, the machine, the case, and judgement — each with learning outcomes, a worked example, graded exercises and further reading. The disciplinary spine is three widely used curricula: COMCEC's introduction to AI and its fundamental concepts, Microsoft's Fundamentals of AI Module 1, and NITI Aayog's Let's Learn AI Base Module, each attributed and mapped chapter by chapter in the appendix. Tapstall runs through the middle of the book as the worked case, including its failures.

Eleven tables, two pseudocode listings, a sixty-term glossary, eight labs and an annotated bibliography.

Contents

1. About this textbook
2. Who should read it
3. How to read it
4. How to use it in teaching
5. A note on evidence and honesty
6. What intelligence is, and why the question is old
7. Four definitions, one discipline
8. A short history that explains today’s defaults
9. What AI can do, what it cannot, and when to use it
10. Intelligence in the small firm
11. Data: the raw material
12. Preparing data: the lifecycle
13. Features: the representation that decides the outcome
14. Seeing data: descriptive statistics and visualisation
15. Learning from data
16. Measuring a model honestly
17. The arithmetic of being wrong
18. Models in production
19. Algorithms, procedures and pseudocode
20. Languages, APIs and the cloud
21. Designing under constraint: the rail
22. The gate: finding prospects without lying to yourself
23. What broke
24. Ethics, consent and the things a system should refuse
25. Law, data protection and the state
26. Teaching it: curriculum, sequence and assessment
27. Conclusion: the discipline of the second question
28. Appendix A · Glossary
29. Appendix B · Eight labs
30. Appendix C · Mapping to the three source curricula
31. Appendix D · References and sources
32. Colophon