SECOND EDITION
Andrew and the Marvellous Analytical Engine
A Practical AI Primer for Lawyers
The second edition is a substantially rewritten practical account of artificial intelligence for lawyers: where the technology came from, what it is now capable of doing, how it can be used in real legal work, and where responsibility must remain with the professional.
A practical book for lawyers who do not intend to become computer scientists.
Artificial intelligence has moved very quickly from novelty to working tool. A lawyer can now use it to explore an unfamiliar subject, analyse documents, organise evidence, draft, compare, calculate and build workflows which only recently would have required substantial human labour. That capability is useful. It is also capable of generating falsehood with considerable fluency, mishandling confidential material and encouraging delegation where supervision was required.
The book takes both propositions seriously. It is neither an argument that AI will replace lawyers nor an exercise in explaining why the profession can safely ignore it.
WHY A SECOND EDITION?
Because twelve months turned out to be a long time.
The first edition was a starting point. During the following year the tools improved markedly, agentic systems became more practical, legal products developed, and the professional consequences became clearer. Courts and regulators also had rather more to say about fabricated authorities, verification, supervision and confidentiality.
The second edition therefore compresses some of the earlier history, removes standalone product chapters and places considerably more weight on practical legal work, professional duties, confidentiality, evidence and the changing economics of practice.
LARGE LANGUAGE MODELS
Useful machinery, without the incense.
The early chapters explain how modern generative AI works in language which assumes intelligence on the part of the reader but no prior knowledge of computer science. They examine large language models, probability, hallucination and the dangerous tendency to confuse a polished answer with an authoritative one.
The practical discipline which follows is simple enough to state: identify the task, control the material where possible, understand the output required, anticipate the relevant failure mode and decide how the result will be checked.
LEGAL WORK
Research, drafting, litigation and agentic workflows.
The central practical chapters consider the uses which matter in day-to-day legal practice. Legal research is treated as a process of discovery, investigation and verification. Drafting is approached through controlled materials and iterative review rather than the hopeful production of a perfect document from one enormous prompt. Litigation chapters deal with chronologies, evidence, document analysis and workflow, while the discussion of agents considers what happens when AI is given tools and permitted to carry out sequences of work.
The machinery may reduce drudgery considerably. It does not acquire responsibility for the document which ultimately reaches a client or a court.
DUTIES AND GOVERNANCE
Regulation, confidentiality and responsibility.
The later chapters turn to professional ethics, regulation, confidentiality, privilege and data governance. Existing professional duties do much of the work. Accuracy still matters. Confidentiality still matters. Supervision remains supervision even where part of the work was generated by software. A lawyer who adopts an argument needs to understand it well enough to defend it.
That also means that “human in the loop” is not a talisman. The review has to be capable of finding the sort of error which the particular use of AI may create.
INSIDE THE SECOND EDITION
Sixteen chapters, followed by practical appendices.
- The Long Dream of the Machines
- Programming the World
- Artificial Intelligence without the Incense
- Large Language Models and Hallucination
- The AI Toolkit for Lawyers
- Legal AI Platforms
- Vibe Coding and Agentic Workflows
- Legal Research
- Drafting and Review
- Litigation Workflows
- Practice Management, Education and Business Development
- AI Ethics for Lawyers
- Regulation, Caselaw and Professional Duties
- Confidentiality, Privilege and Data Governance
- AI, Evidence and Trust
- The Lawyer Who Remains
The book concludes with an Afterword and appendices containing a Model AI Policy for Chambers, Further Reading and a list of Live Resources as they stood at publication.
HOW THE BOOK USED AI
A book about artificial intelligence which used artificial intelligence.
AI assisted heavily with the machinery of drafting, research organisation, testing and illustration. The choices about argument, structure, tone, source checking and final wording remained mine. That is not a confession hidden in the small print; it is part of the subject of the book itself. The interesting question is not whether a machine touched the work, but what the machine was asked to do and who remained responsible for the result.
“A thoughtful and informative guide to AI that will no doubt prove an invaluable handbook for all lawyers — not just those in the middle-aged bracket.”
Law Society Gazette · Five Stars
“An intelligent book about artificial intelligence & lawyers.”
Civil Litigation Brief
Reviews are of the first edition.
BUY THE SECOND EDITION